When Group Membership Gets Personal

16 Pages·2012·0.27 MB·English

PsychologicalReview ©2012AmericanPsychologicalAssociation
2012,Vol.119,No.3,441–456 0033-295X/12/$12.00 DOI:10.1037/a0028589
When Group Membership Gets Personal: A Theory of Identity Fusion
William B. Swann Jr. Jolanda Jetten
UniversityofTexasatAustin UniversityofQueensland
A´ngel Go´mez Harvey Whitehouse
UniversidadNacionaldeEducacio´naDistancia(UNED) UniversityofOxford
Brock Bastian
UniversityofQueensland
Identityfusionisarelativelyunexploredformofalignmentwithgroupsthatentailsavisceralfeelingof
onenesswiththegroup.Thisfeelingisassociatedwithunusuallyporous,highlypermeablebordersbetween
thepersonalandsocialself.Theseporousbordersencouragepeopletochanneltheirpersonalagencyinto
groupbehavior,raisingthepossibilitythatthepersonalandsocialselfwillcombinesynergisticallytomotivate
pro-groupbehavior.Furthermore,thestrongpersonalaswellassocialidentitiespossessedbyhighlyfused
personscausethemtorecognizeothergroupmembersnotmerelyasmembersofthegroupbutalsoasunique
individuals,promptingthedevelopmentofstrongrelationalaswellascollectivetieswithinthegroup.Inlocal
fusion,peopledeveloprelationaltiestomembersofrelativelysmallgroups(e.g.,familiesorworkteams)with
whomtheyhavepersonalrelationships.Inextendedfusion,peopleprojectrelationaltiesontorelativelylarge
collectivescomposedofmanyindividualswithwhomtheymayhavenopersonalrelationships.Theresearch
literatureindicatesthatmeasuresoffusionareexceptionallystrongpredictorsofextremepro-groupbehavior.
Moreover, fusion effects are amplified by augmenting individual agency, either directly (by increasing
physiological arousal) or indirectly (by activating personal or social identities). The effects of fusion on
pro-group actions are mediated by perceptions of arousal and invulnerability. Possible causes of identity
fusion—ranging from relatively distal, evolutionary, and cultural influences to more proximal, contextual
influences—arediscussed.Finally,implicationsandfuturedirectionsareconsidered.
Keywords:identityfusion,socialidentity,personalidentity,groupidentification,self-verification
IamwhatIambecauseofwhoweallare. Everyday,somepeopledoextraordinarythingsfortheirsocial
—CharacterizationoftheAfricanphilosophyUbuntu groups:Theyrisktheirlivesincombat;theydonatetheirpersonal
byLeymahGbowee fortunes;theyevenblowthemselvesup.Inthisarticle,wepropose
that a common mechanism—identity fusion—underlies each of
I’dactuallythrowmyselfonthehandgrenadeforthem...becauseI
theseextremesacrificesforone’sgroup.Weassumethatidentity
actuallylovemybrothers.Imean,it’sabrotherhood...Anyofthem
woulddoitforme. fusionisauniqueformofalignmentwithagroup,onethatentails
—ReflectionsofanAmericansoldierinAfghanistan a visceral feeling of oneness with the group. This feeling of
(SebastianJunger,War) onenessisassociatedwithincreasedpermeabilityoftheboundary
between the personal and social self. Such elevated permeability
willincreasethelikelihoodthatthefusedperson’spersonaliden-
ThisarticlewaspublishedOnlineFirstMay28,2012. tity will influence his or her group identity and vice versa. This
WilliamB.SwannJr.,DepartmentofPsychology,UniversityofTexas articlefocusesonthesemutualinfluenceprocesses.Morebroadly,
atAustin;JolandaJetten,SchoolofPsychology,UniversityofQueensland, we consider the nature and consequences of identity fusion, its
Brisbane,Queensland,Australia;A´ngelGo´mez,DepartmentofSocialand relationship to other forms of alignment with groups, and the
OrganizationalPsychology,UniversidadNacionaldeEducacio´naDistan- variables that cause it and regulate its expression. To place the
cia(UNED),Madrid,Spain;HarveyWhitehouse,SchoolofAnthropology
identityfusionconstructinhistoricalcontext,webrieflyintroduce
andMuseumEthnography,UniversityofOxford,Oxford,England;Brock
itsintellectualancestors.
Bastian, School of Psychology, University of Queensland, Brisbane,
Queensland,Australia.
This work was funded by National Science Foundation Grant BCS- Intellectual Roots of Identity Fusion Theory: From
1124382 to William B. Swann Jr. and Research Fund Grant PSI2009-
Durkheim to the Social Identity Perspective
07008 from the Spanish Ministry of Science and Innovation to A´ngel
Go´mez, Michael Buhrmester, Matthew L. Brooks, J. Francisco Morales,
Fusion-like constructs have a long history in the behavioral
andWilliamB.SwannJr.Forcommentsonearlierversionsofthisarticle,
sciences. Examples include the following: Emile Durkheim’s
we thank David Buss, Elena Gaviria, Jonathan Lanman, Airong Truffet,
(1893/1964, 1915/1995) concept of “mechanical solidarity” and
MatthewHornsey,andYijieWang.
CorrespondenceconcerningthisarticleshouldbeaddressedtoWilliamB. “collective effervescence,” Victor Turner’s (1969) notion of
Swann Jr., Department of Psychology, University of Texas at Austin, 1 “spontaneouscommunitas”(anintensefeelingoftogethernessand
UniversityStationA8000,Austin,TX78712.E-mail:[email protected] common humanity), Alan Fiske’s (1991) notion of communal
441
442 SWANN,JETTEN,GO´MEZ,WHITEHOUSE,ANDBASTIAN
sharing(seealsoSahlins,1974),andHarveyWhitehouse’s(1995) Third,eventhoughthereareindividualdifferencesintheextent
depictionofthe“imagistic”modeofreligiosity.Inthecontempo- towhichsocialcategoriesorgroupmembershipsaresalient(i.e.,
rarypsychologicalliterature,fusion’sclosestintellectualcousinis perceiverreadiness;seeJ.C.Turneretal.,1987),identificationas
“groupidentification”(Lewin,1948). agroupmemberispresumablyregulatedbythesocialcontext.As
Researcherstypicallyunderstandidentificationtoreflectpeo- aresult,changesinthecontextcouldproducesubstantialchanges
ple’s feelings of allegiance to the collective (for other concep- in levels of identification (J. C. Turner, 1999; J. C. Turner et al.,
tualizations,seeBrewer&Gardner,1996;Postmes,Haslam,& 1994).
Swaab,2005;Prentice,Miller,&Lightdale,1994).Thisunder- In recent years, theorists have revisited some of the foregoing
standing is derived from what is known as the “social identity assumptions. Whereas some have questioned the generality or
perspective,” which encompasses social identity theory (e.g., interpretation of the principle of functional antagonism (e.g.,
Tajfel & Turner, 1979) and its extension, self-categorization Abrams,1994;Baray,Postmes,&Jetten,2009;Pickett,Silver,&
theory (e.g., J. C. Turner, Oakes, Haslam, & McGarty, 1994). Brewer, 2002; Postmes & Jetten, 2006; Reid & Deaux, 1996;
Over the last three decades, the social identity perspective has Stephenson, 1981; see also J. C. Turner, Reynolds, Haslam, &
shaped almost all major theorizing regarding group processes Veenstra, 2006), others have contested the depersonalization hy-
(forrecentexamples,seeEllemers,DeGilder,&Haslam,2004; pothesis (e.g., Deaux, 1993; Simon, 2004; Spears, 2001). More-
Ellemers, Spears, & Doosje, 2002; S. A. Haslam, Reicher, & over, although all theorists acknowledge that allegiance to the
collective provides the basis for identification, some have con-
Platow, 2011; Hornsey & Jetten, 2004; Mackie, Devos, &
tendedthatthepersonalrelationshipsofgroupmembersmayalso
Smith,2000;Mackie,Smith,&Ray,2008;E.R.Smith,Seger,
promoteidentification(e.g.,Brewer&Gardner,1996;Postmeset
& Mackie, 2007; Spears, Oakes, Ellemers, & Haslam, 1997).
al., 2005; Prentice et al., 1994). Still others have challenged the
At the heart of the social identity perspective lays the dis-
notionthatidentificationtypicallyfluctuatesmarkedlyinresponse
tinctionbetweenthepersonalandsocialself(e.g.,James,1890;
tocontextualchanges,suggestinginsteadthatitmaybetemporally
Tajfel & Turner, 1979). Whereas the personal self refers to
stable among some individuals (e.g., Roccas, Sagiv, Schwartz,
idiosyncratic properties of the individual (e.g., “intelligent,”
Halevy,&Eidelson,2008).
“sociable”), the social self refers to those aspects of self asso-
Fusiontheoryintegratesalloftheserevisionistthemesbycon-
ciatedwithgroupmembership(e.g.,“Democrat,”“American”).
sideringnewwaysinwhichthepersonalandsocialselvesofgroup
This distinction gives rise to several assumptions, three of
members may relate to one another. In particular, whereas most
which are especially relevant here. First, social identity theory
people experience clearly demarcated boundaries between their
proposes that all interactions with others are located on an
personalandsocialselves,thosewhoarehighlyfusedwithagroup
interpersonal–intergroup continuum (Tajfel & Turner, 1979).
experience this boundary as porous and permeable. We consider
Thisimpliesthatinsofarasgroup-relatedbehaviorismotivated
theimplicationsoftheseporousboundariesnext.
by a salient social self, it will not be motivated by a salient
personalself,andviceversa.Forexample,whenanindividual’s
Nature of Identity Fusion
social identity as a saleswoman is salient and influential, her
personal identity as shy will recede and be less influential.
Identityfusionoccurswhenpeopleexperienceavisceralfeeling
Self-categorizationtheory(J.C.Turner,Hogg,Oakes,Reicher,
of oneness with a group. The union with the group is so strong
& Wetherell, 1987) expanded upon this idea by hypothesizing
among highly fused persons that the boundaries that ordinarily
a hydraulic relationship between the social and personal self—
demarcatethepersonalandsocialselfbecomehighlypermeable.
the “functional antagonism principle.” Social identity ap-
Infact,theseboundariesbecomesopermeablethataspectsofboth
proaches therefore suggest that the motivational burden for
the personal and social self can readily flow into the other. The
pro-group behavior rests entirely on the shoulders of the rele-
flowofinfluencemaymoveinbothdirections:Justashighlyfused
vant social identity (see Hopkins et al., 2007; Levine &
personscometoviewthemselvesthroughtheirgroupmembership
Crowther, 2008).
(“MygroupmembershipisacrucialpartofwhoIam”),theyalso
Second, when social identity is salient and group members perceivethegroupthroughtheirpersonalself(“Iamanimportant
define themselves in terms of their social identity, they will “de- partofthegroup”).Thesemutualinfluenceprocessesencouragea
personalize”—thatis,perceivethemselvesandothergroupmem- strong sense of connection to the group, a sense that motivates
bersascategoricallyinterchangeable.Thismeansthatgroupmem- highlyfusedpersonstodoasmuchforthegroupastheywoulddo
bers will perceive other group members through the lens of their forthemselves.Nevertheless,suchmutualinfluenceprocessesdo
membershipinthegroupratherthanintermsofpersonalrelation- not necessarily diminish the integrity of either the personal or
ships that they have established with one another. Further, their social self. Instead, just as heavy commerce between trading
attraction to fellow group members will be determined by the partnersdoesnotnecessarilydiminishtheintegrityofeithercoun-
degree to which such members embody the prototypic character- try, the connections between the personal and social self do not
isticsofthegroup(i.e.,depersonalizedattraction)ratherthantheir undermine—andmayactuallyincrease—thestrengthandviability
possession of appealing personal qualities (i.e., interpersonal at- ofbothconstructs.
traction;Hogg,1993).Considerableevidencesuggeststhatdeper- The tendency for highly fused persons to maintain permeable
sonalized attraction rather than interpersonal attraction fosters borders between their personal and social selves will not only
group identification (Hogg & Hardie, 1991) and perceptions of magnify their feelings of connectedness to the group category, it
group cohesiveness (Hogg, Cooper-Shaw, & Holzworth, 1993; will also foster connections to other ingroup members. That is,
Hogg&Hains,1996). highly fused persons may project their own strong personal and
IDENTITYFUSION 443
social selves onto other group members. This will introduce the result will be that highly fused persons are unusually willing to
possibility of attraction to other group members based on the engageinpro-groupbehavior.
uniquepersonalselvesofthesemembersaswellastheirmember- 3. Relational ties principle. For highly identified individuals
shipinthegroup.Groupmembersmayaccordinglydevelopstrong whose group memberships are salient, other group members are
relationaltiestoothergroupmembers.Withsuchstrongrelational mere carriers of information regarding what the group stands for
ties will come a sense of obligation to help and defend fellow (e.g.,thenormsofthegroup).Moreover,attractiontoothergroup
groupmembers.Moreover,justashighlyfusedpersonswillbe members is based only on the degree to which group members
inclinedtobelievethattheythemselveswilldoanythingforthe embodytheprototypicqualitiesofthegroup.Incontrast,asnoted
group and its members, they will project their feelings of above, highly fused persons possess strong personal as well as
personal agency onto others. As a result, they will develop the socialidentitiesandtendtoassumethatothersdoaswell.Highly
conviction that other group members are similarly disposed to fused persons will thus be predisposed to recognize the unique
protect the group and its individual members. This assumption personal identities as well as social identities of fellow group
thatothergroupmembersarealsoextraordinarilycommittedto members, allowing for “uniqueness-based” as well as
thegroupanditsindividualmemberswillfostertheperception “membership-based” attraction (akin to Hogg’s interpersonal and
of reciprocal strength, wherein highly fused individuals will depersonalized attraction, respectively; see Hogg, 1993; Hogg et
perceive that the group is not only extremely powerful, but al., 1993; Hogg & Hardie, 1991).1 The combination of
invulnerable due to the combined effect of personal and group membership-based and uniqueness-based attraction may produce
agency. exceptionally strong relational ties: Other ingroup members will
Inshort,thestateofidentityfusionreferstoapowerfulunionof not only be valued by virtue of their representativeness of the
the personal and social self wherein the borders between the two group(prototypicalityforthegroup),theywillalsobevaluedfor
become porous without diminishing the integrity of either con- their unique personal characteristics that make them attractive
struct.Theresultisapowerfulfeelingofconnectedness,notonly relationshippartners.
tothegroupcategorybutalsototheothermembersofthegroup. The strong relational ties associated with fusion may play out
These feelings of connectedness, in turn, foster strong relational differentlydependingonwhetherfusionislocalversusextended.
ties to other group members and the perception of reciprocal In local fusion, group members form relational ties with others
strength.Thisreasoninggivesrisetothefourprinciplesofidentity with whom they have direct personal contact and thus have the
fusiondiscussedbelow.Theseprinciplescapturethemostimpor- opportunitytoshareexperiences.Thiscommonlyoccursintribal
tantwaysthatfusionisdistinctfrompreviousformsofalignment units, small bands of teammates or soldiers, and other close-knit
withgroups,suchasidentification. groups. In extended fusion, people may project the relational ties
1. Agentic-personal-self principle. One distinction between normally associated with local fusion onto large groups despite
identificationandfusioninvolvestheroleoftheagentic-personal- havinglittleornodirectcontactorsharedexperienceswithindi-
self in pro-group behavior. The social identity perspective holds vidualmembers.Duetothelackofpersonalcontactwithallother
thatwhenasocialidentityissalient,theactionsofhighlyidenti- groupmembers,the“relationalties”thatpeopledevelopwithother
fiedpersonsareregulatedbya“depersonalized”socialselfasso- groupmembersaremetaphoricalratherthanliteral.Membersofa
ciated with the group; feelings of personal agency presumably given ethnicity, for example, may fuse with ethnically similar
playnoroleinpro-groupactivities.Incontrast,whenpeoplefuse members of nation states (e.g., Spain, China) based on common
withagroup,theydonottemporarilyabdicatetheirpersonalself. ancestry,despitehavingfewornosharedexperienceswithallof
Rather, when highly fused persons enact pro-group activity, them.Alternatively,peoplemayfusewithgroupsonthebasisof
their actions reflect both their personal and social identities, anabstraction,suchasacommoncauseorimportantvalue.What-
working together by virtue of the porous borders that define ever form fusion takes, the relational ties principle raises the
them. This motivational dynamic increases the chances that possibility that the intra-psychic synergies of individual highly
suchindividualswilltethertheirfeelingsofpersonalagencyto fusedpersonsmaycombinewithinterpersonalsynergieswherein
the group’s agendas. It will also introduce the possibility of highly fused persons encourage other group members to enact
synergistic relations between the personal and social selves of pro-groupbehavior.Theresultofsuchencouragementmaybethe
highly fused persons. enactmentofunusuallyboldandpotentiallydangerousactionson
2. Identity synergy principle. Because group behavior is pre- behalfofthegroup.
sumably fueled exclusively by social selves (e.g., J. C. Turner et 4. Irrevocability principle. Highly identified persons should
al., 1987), personal selves contribute minimally to pro-group be- remain devoted to the group only insofar as the immediate con-
havior. In contrast, the identity synergy principle holds that the textual influences support such devotion; removal of contextual
personalandsocialidentitiesofhighlyfusedpersonsmaycombine support for identification may therefore produce corresponding
synergisticallytomotivatepro-groupbehavior,therebyproducing
additional motivational “oomph.” The identity synergy principle
1Intheinterestoftheoreticalprecision,wehavereplacedHogg’sterms
also suggests that it should be possible to amplify pro-group
“depersonalized”and“interpersonal”attraction,with“uniqueness-based”
behavior among highly fused persons by activating either their
and“membership-based,”respectively.Basedonfusiontheorizing,valuing
personalorsocialself-views.Thatis,duetotheporousboundaries
groupmembersbecausetheyaregroupmembersdoesnotimplythatwe
betweenthepersonalandsocialselfofsuchindividuals,activating
perceive them as less “person-like” (i.e., depersonalized), and valuing
eitheronewillactivatetheother,therebypromotingactivitiesthat othersfortheiruniquequalitiesbydefinitionisnotmore“interpersonal”
areemblematicoftheindividual’scommitmenttothegroup.The thanvaluingthemasagroupmember.
444 SWANN,JETTEN,GO´MEZ,WHITEHOUSE,ANDBASTIAN
diminutionsinidentification.Bycontrast,oncefused,peoplewill other” (Aron et al., 1992, p. 598). Several group researchers
tend to remain fused. One reason for this is the relational ties (Coats, Smith, Claypool, & Banner, 2000; E. R. Smith & Henry,
principle,whichintroducesthepossibilitythatactualorimagined 1996; Tropp & Wright, 2001) adapted the IOS to capture align-
relationalbondstoothergroupmembers—inadditiontodevotion mentofrespondentswithgroups.Buildingonthiswork,Schubert
to the collective—will buttress feelings of fusion. For example, andOtten(2002)addedanoptioninwhichtheselfandgroupwere
fused people may develop close relationships with other group completelyoverlapping.Swannetal.(2009)furthermodifiedthe
memberswhosupportandthusstabilizetheirfeelingsofalignment pictorialmeasurebyaskingrespondentstoindicatewhichoffive
withthegroup.Furthermore,theexclusivityoffusion(i.e.,people representations of the self and group best captured their relation-
tend to fuse with only one group within a given group category, ship with the group. Importantly, in the pictorial measure of
such as nation or religion; Swann, Go´mez, Seyle, Morales, & identityfusiondisplayedinthetoppanelofFigure1,thepictorial
Huici,2009)willencouragefusedpeopletosever,orrefrainfrom representation of the personal self remained visible despite being
cultivating, alignments with other groups, thereby diminishing completelymergedwiththegroup.
their alternatives to the fused group. In this way, their powerful Scores on the pictorial measure of fusion were distributed bi-
alignment with the group will lock highly fused persons into modally,with“fused”personsselectingthemostextremeoptionin
self-perpetuating interpersonal cycles that stabilize the very psy- whichthecirclerepresentingthe“self”wascompletelyimmersed
chological structures that drew them into their groups in the first in the larger circle representing the “group.” Non-fused persons
place. selected the other four options. The tendency for people to fuse
Finally, the porous borders between the personal and social with a particular group was fairly stable over time (r (cid:1) .56;
selves of highly fused individuals will mean that both types of Go´mez,Brooks,etal.,2011).Nevertheless,fusionresemblesan
identities support the allegiance that highly fused persons feel attituderatherthanatraitinthatthetendencytofusewithone
toward the group. Such allegiance may be particularly important group(e.g.,one’sreligion)wasuncorrelatedwiththetendency
when people encounter some threat to their personal or social tofusewithothergroups(e.g.,one’scountry;allrs(cid:2).11,ns).
identities.Thatis,aswenotebelow(e.g.,Go´mez,Morales,Hart, The group-specificity of fusion presumably reflects a tendency
Va´zquez,&Swann,2011;Swannetal.,2009),whenhighlyfused for significant emotional commitments to a given group to
individuals encounter challenges to their personal or social iden- preclude equally significant commitments to rival groups.
tities,theirdesireforstableself-viewsmaytriggercompensatory Inoveradozenstudies,thepictorialmeasureoffusionsuccess-
self-verificationstrivings(Swann,2011).Thesecompensatoryac- fullypredictedendorsementofpro-groupbehaviorssuchasfight-
tivities will reaffirm the identities that have been challenged and inganddyingforone’scountry(e.g.,Go´mez,Brooks,etal.,2011;
thus shore up the feelings of fusion that they support. Together Swannetal.,2009).Moreover,inseveralvariationsoftheclassic
with the relational ties and exclusivity tendencies noted above, trolley dilemma, highly fused persons (but not non-highly fused
thesecompensatoryactivitiesmayencouragepeoplewhobecome persons) endorsed saving fellow group members by plunging
fusedtoremainfused. themselves in front of a speeding locomotive (Swann, Go´mez,
Dovidio, Hart, & Jetten, 2010). All of these findings emerged
Empirical Research on Identity Fusion while controlling for group identification. Furthermore, Swann,
Go´mez,Huici,Morales,andHixon(2010)reportedthatalthough
Severalstudieshavebeenconductedtotesttheviabilityofthe fusionwascloselyassociated(B(cid:1)0.53,p(cid:1).001)withameasure
fourprinciplesoffusiontheory.Tosetthestageforthesestudies, of commitment (Rusbult & Farrell, 1983), fusion was a stronger
researchers developed and validated two measures of fusion. We predictor of endorsement of extreme actions for Spain than com-
summarize this validational work prior to describing tests of the mitment, z (cid:1) 5.99, p (cid:1) .001, Bs (cid:1) 0.71 and 0.36, respectively,
fourprinciples. ps(cid:1).001.
Yet,ifthestrongtrackrecordofthepictorialmeasureoffusion
Developing and Validating Measures of Identity supportedtheadage“apictureisworthathousandwords,”itwas
unfortunatelyunclearpreciselywhich“words”participantshadin
Fusion
mindwhentheyendorsedthefusedoption.Toaddressthisshort-
The goal of this work was to develop instruments that could coming,Go´mez,Brooks,etal.(2011)developedandvalidatedthe
identify participants who felt so strongly unified with a social verbalmeasureofidentityfusiondisplayedinthebottompanelof
groupthattheboundariesbetweentheirpersonalandsocialselves Figure 1. They began by assuming that the porous boundaries
werehighlypermeable.Toaccomplishthisaim,researchersbegan between the personal and social selves among highly fused
by modifying an existing pictorial methodology that had been persons would give rise to two complementary aspects of fu-
developedtoassessattachmentincloserelationships:theInclusion sion.Thefirstaspectoffusioninvolvedfeelingsofconnected-
of Other in the Self Scale (IOS; Aron, Aron, & Smollan, 1992; ness or “oneness” with the group. To assess these feelings of
Aron, Aron, Tudor, & Nelson, 1991; see also Cialdini, Brown, connectedness,theresearchersincludeditemssuchasthe“Iam
Lewis,Luce,&Neuberg,1997;Maneretal.,2002).Comprisedof onewithmycountry,”and“Ifeelimmersedinmycountry”(see
a series of pictures that represent different degrees of overlap the bottom panel of Figure 1). The second aspect of fusion
between the self and other, the IOS was conceptualized as a involvedtheperceptionthatthegroupmembermakesthegroup
measureofclosenesstoanotherperson,thedegreetowhichpeople strongandthegroupmakesthepersonstrong.Tomeasurethis
possess a “sense of being interconnected with another.” Such reciprocal strength aspect of identity fusion, the researchers
interconnectednesstheoreticallyentailsatendencytoviewtheself included items such as “I am strong because of my country,”
as “including resources, perspectives, and characteristics of the and “I make my country strong.”
IDENTITYFUSION 445
Figure 1. Measures of fusion pictorial (top panel; Swann et al., 2009) and verbal (bottom panel; Go´mez,
Brooks,etal.,2011).
AnalysisofEnglishandSpanishlanguageversionsofthe7-item Ashforth’s(1992)scalerevealedtwofactors,withthefirstfactor
fusionscaleindicatedthatithaddesirablepsychometricproperties. includingthesevenitemsfromtheverbalmeasureoffusionanda
Forexample,afactoranalysisrevealedthatallitemsloadedona secondfactorincludingthesixitemsfromtheidentificationscale
single factor and that the coefficient (cid:3) was .84. The test–retest (Go´mez,Brooks,etal.,2011,p.922).Confirmatoryfactoranalysis
correlationovera6-monthperiodwasrespectable,r(618)(cid:1).71, based on an independent sample verified the two-factor solution
p(cid:2).001,exceedingthestabilityofthepictorialmeasureoffusion, (Go´mez,Brooks,etal.,2011,Figure1,p.923).
r(618)(cid:1).56,p(cid:2).001,z(cid:1)4.16,p(cid:2).001—inturn,thetest–retest Tofurtherassesstheconvergentanddiscriminantvalidityofthe
stabilityofthepictorialfusionscaleexceededthestabilityofMael verbal measure of fusion, Go´mez, Brooks, et al. (2011) related
andAshforth’s(1992)identificationscale,r(618)(cid:1).44,z(cid:1)2.82, scoresonthescaletoscoresonseveralotherscalesthatwere,or
p(cid:2).01.Moreimportant,scoresontheverbalmeasureoffusion were not, expected to be related to fusion. As noted above, the
were closely associated with scores on the pictorial measure of verbalmeasureoffusionwasstronglycorrelatedwiththepictorial
fusion,withdisattenuatedcorrelationsrangingfrom.87to1.0.2 measureoffusionwhenthecorrelationsweredisattenuated.Atthe
Go´mez, Brooks, et al. (2011) also assessed the discriminant same time, fusion was unrelated to several potentially related
validityoftheverbalmeasureofidentityfusion.Forexample,they personalityattributes.Analysesrevealedthatscoresontheverbal
tested the hypothesis that fusion is related to, but distinct from measure of fusion were unrelated to individual differences in
group identification, the standard measure of alignment with self-concept clarity, empathy, or aggressiveness. Weak relations
groups. As an index of identification, the researchers focused on emerged between fusion and both self-efficacy and essentialism
MaelandAshforth’s(1992)well-respectedandwidelycitedscale. (rs(cid:2).18).Thesefindingsareimportantbecausetheysuggestthat
There were two reasons for choosing this scale. First, earlier the substantial temporal stability of scores on the verbal measure
research (Swann et al., 2009) indicated that Mael and Ashforth’s offusiondoesnotreflectatendencyforittomasqueradeasoneof
scalewasmorestronglyassociatedwithfusion,r(198)(cid:1).56,p(cid:2) theforegoingtraits.
.01, than either of two rival scales: Jetten, Branscombe, Schmitt, Havingestablishedtheconvergentanddiscriminantvalidityof
andSpears’s(2001)scale,r(112)(cid:1).26,andTroppandWright’s the verbal measure of fusion, Go´mez, Brooks, et al. (2011) next
(2001) scale, r(248) (cid:1) .23. Second, in head-to-head competition conductedaseriesoftestsofitspredictivevalidity.Considerable
with Leach et al.’s (2008) recently developed scale, Mael and
Ashforth’s scale was the stronger predictor of endorsement of
extremepro-groupbehavior(Swann,Go´mez,Huici,etal.,2010). 2Correlationsbetweentheverbalandpictorialmeasuresoffusionare
availablefromsixsamples(threedifferentsubjectpopulationscompleted
This evidence suggests that of extant identification scales, Mael
themeasuresseveralmonthsapart).Thecorrelations(rangingfrom.52to
and Ashforth’s scale is the one that predicts extreme pro-group
.67)wereattenuatedbythefactthatoneofthecovariates(i.e.,thepictorial
behavior with the most fidelity, making it the most appropriate
measure)wasasingle-itemscale,andsuchscalesareknowntobenoto-
standardofcomparisonwiththefusionscale.
riouslyunreliable.Consistentwiththispossibility,whenwecorrectedthe
Psychometric analyses supported the distinction between the correlationbetweenthetwoscalesforattenuation(usingtest–retestcoef-
fusion and identification scales. For example, exploratory factor ficientsasourindexofreliability),thedisattenuatedvaluesrangedfrom
analysis of all items from the identity fusion scale and Mael and .87to1.0.
446 SWANN,JETTEN,GO´MEZ,WHITEHOUSE,ANDBASTIAN
support emerged. For example, when scores on the fusion scale compensatoryeffortstoreaffirmtheseself-views.Thus,forexam-
wereusedtopredictendorsementoffightinganddyingforone’s ple, when people learn that others do not see them as they see
group 6 months later, the verbal measure of fusion predicted the themselves, they compensate by working to bring that person to
outcomevariableswithgreaterfidelitythanthepictorialmeasure viewthemastheyviewthemselves—evenwhenthisentailslow-
offusionaswellasameasureofidentification(Go´mez,Brooks,et eringaninteractionpartners’overlypositiveevaluationofaneg-
al.,2011,Study5andStudy7b).Inaddition,theverbalmeasure ativequality(e.g.,Brooks,Swann,&Mehta,2011;Swann&Hill,
of fusion outstripped its rivals in predicting the likelihood that 1982;Swann&Read,1981;Swann,Wenzlaff,&Tafarodi,1992).
participants would endorse jumping to their deaths in front of a Ifchallengestopersonalself-viewstriggercompensatoryreac-
speedingtrolleytosaveafellowgroupmember(Study6a)orkill tions, and the boundaries between the personal and social self-
terrorists who threatened the group (Study 6b). Finally, whereas views of highly fused persons are highly permeable, then chal-
the foregoing studies were conducted with Spaniards, additional lenges to either their personal or social identities should amplify
studiesvalidatedtheverbalmeasureoffusionwithtwosamplesof pro-groupbehavior.Totestthishypothesis,researchersfirstchal-
immigrantsfrom22differentnations(Studies7aand7b)aswell lenged the personal self-views of fused and non-highly fused
as a sample of Americans who completed an English language persons by providing them with unexpectedly positive feedback
versionofthescale(Study8). about personal qualities (of course, providing participants with
These findings therefore led to two important conclusions. unexpectedlynegativefeedbackshouldalsotriggercompensatory
First, the verbal measure of fusion was closely associated with activity, but such activities could reflect either self-verifying or
the pictorial measure of fusion. Second, on all indices of self-enhancementmotivations,anambiguitythattheinvestigators
construct validity, the verbal measure of fusion exceeded the soughttoavoid).Afterthechallengemanipulation,theresearchers
pictorial measure of fusion, and the pictorial measure, in turn, assessed the degree to which participants endorsed pro-group
exceeded the identification scale. These findings set the stage behaviors. As predicted, challenging participants’ personal self-
for a program of research designed to determine whether indi- views increased subsequent endorsement of pro-group activity
ces of fusion behaved in accord with the four principles of among highly fused persons but not among non-highly fused
fusion theory. persons(Swannetal.,2009,Experiments1and2).
Apparently, the desire of highly fused persons (but not non-
highly fused persons) to compensate for a challenge to their
The Agentic-Personal-Self principle
personalself-viewssynergisticallyamplifiedtheirendorsementof
Insofar as highly fused individuals are poised to channel their pro-group action. A complementary test of the identity synergy
feelingsofagencyintopro-groupbehavior,anymanipulationthat principle turned on independently activating personal and social
increasesagencyshouldtheoreticallyincreasepro-groupbehavior. identities of highly fused persons by first asking them how they
One of the most direct ways to increase agency is to increase would respond in the wake of a physical attack on themselves (a
autonomic arousal through various activities such as physical challenge to the personal self) or an attack on their group (a
exercise(Jacobs&Farel,1971). challenge to the social self). Later, the investigators assessed
To test the proposed link between autonomic arousal and pro- participants’ endorsement of pro-group actions. The results indi-
groupactivityamonghighlyfusedpersons,researchersconducted cated that activating either the personal or social identities of
fourexperimentsinwhichtheyexperimentallyinducedphysiolog- highly fused persons increased their subsequent propensity to
icalarousalbyhavingparticipantseitheroperateanexer-cycle,run endorsefightingordyingforthegroup.Nosuchpatternemerged
windsprints,orplaydodgeball.Asexpected,increasesinarousal amongnon-highlyfusedpersons(Swannetal.,2009,Experiment
amplified the tendency for highly fused persons to translate their 3;seealsoBarayetal.,2009).
feelingsoffusionintopro-groupbehavior,includingovertbehav- Replicatingtheearlierfindings,theforegoingfindingsindicated
iors such as motor activity (i.e., racing a group-related avatar) or thattheflowofactivationcanmovefromthepersonaltothesocial
donating personal funds to a needy group member. Moreover, self. One additional finding indicated that the flow of activation
perceptions of agency mediated the interactive effects of arousal canmoveintheoppositedirection,fromthesocialtothepersonal
and fusion on pro-group behavior (Swann, Go´mez, Huici, et al., self.Thatis,activatingthesocialselvesoffusedparticipants(but
2010). An independent investigation replicated this meditational not non-fused participants) increased subsequent ratings of the
evidencebyconfirmingthatagencymediatedtheimpactoffusion certaintyofpersonalselves(Swannetal.,2009,Study3).Finally,
on endorsement of pro-group behavior (Go´mez, Brooks, et al., evidenceofcompensatorypro-groupactivityfollowingchallenges
2011). topersonalself-viewswasreplicatedinalaterinvestigation(Go´-
mez,Brooks,etal.,2011,Study10)inwhichtheverbalmeasure
of fusion served as a predictor of endorsement of extreme pro-
Identity Synergy Principle
groupbehavior.
One set of tests of the identity synergy principle drew upon a Arecentseriesofstudiesoftherelationshipofidentityfusionto
paradigm developed by researchers interested in self-verification reactions to irrevocable social ostracism provides further support
theory(Swann,1983,2011).Thetheoryassumesthatpeoplehave for the identity synergy principle. In these studies, ostracizing
adeepneedforsocialrelationsthatarecoherent,predictable,and participants for either their personal preferences (i.e., personal
devoid of conflict and misunderstanding. They accordingly de- self-views) or their group membership (i.e., social self-views)
velopapreferenceforotherstoseethemastheyseethemselves, amplified their subsequent endorsement of three distinct types of
eveniftheirself-viewshappentobenegative.Moreover,feedback compensatory activities: endorsement of extreme actions for the
that challenges their negative (or positive) self-views will trigger group, stiffened resolve to remain in the group, and increased
IDENTITYFUSION 447
charitabledonationstothegroup(Go´mez,Morales,etal.,2011).
Hence, as in the initial studies (Swann et al., 2009), the findings
supported the identity synergy principle by showing that it was
possible to amplify pro-group action by activating either the per-
sonal or social identities of fused participants. Apparently, for
highly fused persons, the boundaries between their personal and
socialselvesaresufficientlyporousthatbeingexcludedfromthe
groupissoupsettingthatittriggersself-sacrificialbehavior.This
pattern, however, did not emerge among non-fused participants,
suggesting that the boundaries between their personal and social
self-viewswererelativelyimpermeable.
Relational Ties Principle
The relational ties principle suggests that highly fused persons
Figure 2. Invulnerability and agency mediate the effect of fusion on
should be especially inclined to indicate that they are willing to endorsementofextremebehaviorforthegroup.
sacrificetheirlivestosavethelivesoffellowgroupmembers.To
testthisproposition,Swann,Go´mez,Dovidio,etal.(2010)created
several interpersonal variations of the classic trolley dilemma. In
onestudy,participantswhowerefusedwiththeircountryendorsed thegroup.Thesefindingsreplicatedandextendedearlierevidence
savingfellowSpaniardsbyjumpingtotheirdeathsinfrontofthe that agency mediates the impact of fusion on pro-group behavior
speeding trolley. Two additional studies showed that the self- (Swann,Go´mez,Huici,etal.,2010).
sacrificial behaviors of fused participants generalized to saving
membersofan“extendedfamily”(Europeans)butnotmembersof Irrevocability Principle
an outgroup (Americans).3 In a final study, fused participants
endorsedpushingasideafellowSpaniardwhowaspoisedtojump Researchershavetestedthe“oncefused,alwaysfused”hypoth-
tohisdeathandjumpingthemselves,therebyinitiatingachainof esis by comparing the temporal stability of fusion-with-country
eventsthatwouldostensiblyleadtothedeathsofseveralterrorists. scoresofSpaniardswhowerehighlyfusedwiththoseofpersons
Apparently,highlyfusedpersonsaresostronglyalignedwiththeir who were only moderately or weakly fused. In four independent
fellowgroupmembersthattheywouldpreferthattheythemselves, samples (Ns (cid:1) 513, 155, 421, 219) collected over delays
rather than a group member, should die. In all four studies, ranging from 1 to 18 months, strong test–retest correlations on
participantswhowerenothighlyfusedwerereluctanttosacrifice the verbal measure of fusion emerged among those who were
themselves. highly fused (i.e., those who scored in the upper tertile on
The desire of highly fused persons to help fellow group mem- fusion initially; rs (cid:1) .61, .59, .62, .54). The stability coeffi-
bers has also been shown to influence overt behavior. That is, cients for highly fused participants were significantly higher
highly fused persons were especially inclined to donate their (all Zs (cid:4) 1.88, ps (cid:2) .03) than the coefficients associated with
personalfundstoaneedymemberoftheingroup.Moreover,this participants from the lower (rs (cid:1) .36, .28, .22, .24) or middle
tendency was amplified when participants were physiologically (rs (cid:1) .21, .35, .33, .23) tertiles.4
aroused(Swann,Go´mez,Huici,etal.,2010). In short, there is some rudimentary evidence for each of the
Asnotedabove,therelationaltiesprinciplesuggeststhathighly principles of identity fusion theory. That is, the research liter-
fusedpersonswillfeelthattheyandothergroupmemberssyner- aturesuggeststhatfusionisastateinwhichthepersonalselfof
gistically strengthen each other. This perception of reciprocal group members remains agentic, can synergistically motivate
strengthshouldfostertheperceptionthattogether,themembersof groupbehavior,involvesrelationalties,andisrelativelystable.
the group are uniquely invulnerable. Feelings of invulnerability Nevertheless, the extant literature has provided relatively little
havebeenlinkedtothepropensitytoengageindangerousbehav- insightintotheunderlyingcausesoffusion.Furthermore,little
ior(e.g.,Greene,Krcmar,Walters,Rubin,&Hale,2000;Ravertet hasbeensaidregardingtheboundaryconditionsoffusion,such
al.,2009).Hence,fusionmayfosterperceptionsofinvulnerability as the conditions under which the process of fusion will be
(aswellasagency),andsuchperceptionsshould,inturn,motivate
extreme pro-group behavior. To test the possibility that percep-
tions of invulnerability and agency might mediate the effects of 3Other research has followed up on this finding by showing that al-
fusion on endorsement of pro-group behavior, researchers tested thoughthereissomelimiteddegreeofspilloverfromfusionwithagiven
entity to related entities, there is also specificity. For example, Spanish
themeditationalmodeldisplayedinFigure2(Go´mez,Brooks,et
participantswhowerefusedwithSpainexpressedmorewillingnesstofight
al.,2011,Study9).Theydiscoveredthat(a)fusionwasrelatedto
and die for Spain than for Europe, whereas those fused with Europe
thetwomediators(invulnerabilityandagency),(b)thetwomedi-
expressed more willingness to fight and die for Europe than for Spain
ators were related to endorsement of pro-group action, and (c)
(Swann,Go´mez,Huici,etal.,2010).
controllingfortheeffectsofthemediatorseliminatedtherelation- 4Wefocusontest–retestcorrelationsherebecausethisisthestandard
ship between fusion and endorsement of pro-group behavior. procedure for assessing the stability of the rank-orderings of individuals
Hence, perceptions of invulnerability and agency fully mediated (e.g.,Nunnally&Bernstein,1994).Notably,parallelanalysesusingiden-
therelationshipoffusiontoendorsementofextremebehaviorfor tificationastheindexofalignmentwiththegrouprevealednosuchpattern.
448 SWANN,JETTEN,GO´MEZ,WHITEHOUSE,ANDBASTIAN
reversed and the person will “defuse” from the group.5 We priortoadulthood(Lieberman,Tooby,&Cosmides,2007).Insofaras
consider these issues next. long-termassociationpromotesfusionandfusionpromotesreciprocal
cooperationandaltruism,fusioncouldhaveevolvedasamechanism
Causes of Identity Fusion fordemarcatingtribalgroupsandmaximizingtheinclusivefitnessof
individualswithinsuchgroups.
Thewillingnessofpeopletoengageinextremebehaviorssuch Nevertheless,theprincipleofinclusivefitnesscannotexplainthe
asfightinganddyingforone’sgroupisbafflingbecauseitappears willingnessofhumanstofusewithlargegroupsofgeneticallyunre-
tobeatoddswitharational,cost-benefitanalysis.Here,wehave lated individuals (i.e., extended fusion). For instance, modern-day
attempted to solve this puzzle by focusing on one potential root suicidebombersandmembersofmodernmilitaryunitsappeartofuse
causeofself-sacrifice,namelyidentityfusion.Thisshiftraisesan withpersonswhoaretoodistantlyrelatedgeneticallyforHamilton’s
entirely new set of questions. For example, what are the genetic
(1964a,1964b)ruletoapply.Onewaytorescueaninclusivefitness
andculturalfactorsthatencouragethedevelopmentoffusion?
explanationfortheevolutionoffusionistoassumethatfusionwasfar
Webeginbyofferinganevolutionaryperspectiveonthecauses
morelikelytoberestrictedtosmalltribalgroupsintheancestralpast
of fusion and then discuss various more proximal, cultural influ-
thanistypicallythecasetoday(forotherexplanatoryframeworks,see
encesonfusion.Notably,distalandproximalcausesshedlighton
Buss, 2012; Nowak, 2006; Tooby & Cosmides, 2010). In this sce-
different aspects of the fusion process. For example, the evolu-
nario,initially,fusionemergedasabiologicaladaptationsupporting
tionary perspective speaks mostly to the origins of relational ties
cooperation within tribal groups. Later, however, it served increas-
andhowthesetiescanbeprojectedontogroupswherenogenetic
ingly as a means of solving collective action problems on a much
relationships exist. Consideration of more proximal causes and
largerscale.Ineffect,thinkingintermsoftribalunits(andthefusion
influenceshelptounderstandtheconditionsthatfostertheemer-
itengendered)waseventuallyextendedtolargercommunities.Oneof
genceoffeelingsofconnectedness,reciprocalstrength,andshared
thehallmarksofsuchgroupsisthatmembersfeeltheysharesome-
categorymembershipthatarethehallmarksoffusion.
thingincommon,suchasblood,deepattraction,nationalidentity,or
Evolutionary origins of fusion. Both Darwin (1859) and
a history of suffering. In this way, the tendency to fuse with other
Wallace (1870; co-founder of evolutionary theory) struggled to ex-
groupmembersmayhavegraduallyshiftedfromlocalfusion(often
plainself-sacrificeforthegroupwithinanaturalselectionframework.
based on genetic relatedness) to something resembling the modern
Tothisend,theyaddedscientificallyquestionableparameterssuchas
constructof“sharedessence,”asenseofdeep,underlyingsimilarity
“moral virtue” or “spiritual essence” to the “survival of the fittest”
mechanism in evolutionary theory. Roughly a century after Darwin thatprovidesabasisforextendedfusion(Medin&Ortony,1989).
wrestledwiththeparadoxofhumanself-sacrifice,Hamilton(1964a, Within a shared essence framework, “ingroups” and “out-
1964b) introduced the notion of inclusive fitness. He reasoned that groups”areunderstoodtoresemblenaturalkindsorspecies(Roth-
becausecloserelativesofanorganismarelikelytosharegeneswith bart & Taylor, 1992), especially when they are highly entitative
thatorganism,behaviorsthatpromotethesurvivalofsuchrelatives (McGarty,Haslam,Hutchinson,&Grace,1995;Yzerbyt,Rocher,
willincreasethelikelihoodthatthesharedgeneswillbepassedon. & Schadron, 1997) or organized around endogamy and descent
Thiscouldexplainwhyindividualsmightcompromisetheirindivid- (Gil-White,2001).Indeed,peopleseemtohaveanaturalpropen-
ualfitnesstobolstertheir“inclusivefitness”(whereinfitnessincludes sitytoparsethesocialworldinwaysakintotheirunderstanding
genessharedwithrelatives).Hamiltonbuttressedthisconclusionwith of the biological world (Gelman, 2003; Hirschfeld, 1996), moti-
amathematicalproofthatdemonstratedthattheinclusivefitnessofan vating the perception of large social aggregates “as-if” they are
organism varies as a function of the sum of its own reproductive genetically related. Such essentialistic thinking has implications
success(classicfitness)plustheeffectsoftheorganism’sactionson for how people perceive groups (Bastian & Haslam, 2006, 2007)
thereproductivesuccessofgeneticrelatives(weightedbythedegree as well as how they respond to markers of group membership
of relatedness of the relatives). In subsequent years, researchers re- (Bastian & Haslam, 2008; Bastian, Loughnan, & Koval, 2011;
ported studies of diverse taxa, ranging from amoebas (Strassman, Chao, Chen, Roismann, & Hong, 2007; M. J. Williams & Eber-
Zhu,&Queller,2000)toprimates(Buchan,Alberts,Silk,Altmann, hardt,2008).Ofparticularrelevancehere,modernhumansimpute
2003; Chapais & Berman, 2004; Silk, 2002; Tofilski et al., 2008), sharedessencetoentiregroupsofgeneticallyunrelatedindividuals
indicating that willingness to self-sacrifice increases with genetic (Gelman,2003;N.Haslam,Rothschild,&Ernst,2000;Hirschfeld,
relatedness.Hence,fromthevantagepointofevolutionarytheory,the
1996; Medin & Ortony, 1989; Rothbart & Taylor, 1992). Such
crucialoutcomeisthesurvivalofthegeneratherthantheindividual
whohappenstobecarryingthegene:The“fittest”in“survivalofthe
fittest”referstoageneratherthantheindividual(Dawkins,1976). 5Identityfusionrequirestheexistenceofasocialgroupbecause,bydefi-
Thecognitivearchitectureresponsibleforfusionmayhaveevolved nition,itinvolvestheunionofapersonalandsocialidentity.Nevertheless,
undernaturalselectionasamechanismdesignedtodetectthegenetic- individualsmayalsofusewithabstractions.FeelingsofonenesswithGod,for
relatedness of other group members. Clearly, genetic relatedness example,maycompelpeopletodevotetheirlivestothepriesthood.Similarly,
peoplemaydevelopa“calling”—apowerfulurgetopursuesomeprofessional
cannotbeobserveddirectly.Althoughcuessuchasphenotypicsim-
orrecreationalactivity—thatispropelledbyfeelingsoffusion.Further,people
ilarity may be used as evidence of paternity (Alvergne, Faurie, &
mayfeelfusedwithbrandsorproducts.Ineachoftheseinstances,highlyfused
Raymond, 2009; Platek, Burch, Panyavin, Wasserman, & Gallup,
personsmayexperiencefeelingsofonenessandconnectednesswiththetarget
2002),suchcuesarepronetoerrorandarelimitedinscope.Instead,
oftheirdevotioneventhoughthereisnogroupassociatedwiththetargetof
whenhumanslivedinsmallgroups,theymayhaveinferredgenetic theirattachment.Theabsenceofagroupinsuchinstances,however,means
relatednessfromdegreeofassociationwithothergroupmembersor that it is inappropriate to consider these examples of alignment with an
eventhesheeramountoftimetwoindividualsspentwithoneanother abstractiontobeinstancesofidentityfusion.
IDENTITYFUSION 449
attributionsofsharedessencetoothergroupmembersmayallow Note that cultural ideologies and group norms may not only
forandenhancefusionwiththegroup. encourage feelings of fusion, they may also shape the manner in
Inshort,thetendencyforpeopletofusewithgroupsandmake whichsuchfeelingsareexpressed.Thus,forexample,justassome
pro-group sacrifices may be a by-product of one or more basic groups may develop norms that endorse pacifism as a means of
survival mechanisms. These mechanisms may form a platform achieving the goals of the group, other groups may advocate
upon which ideologies and cultural practices may develop that extremebehaviorssuchassuicidebombingsasameansofaccom-
furthercultivateasocialenvironmentthatfostersthedevelopment plishingagreedupongoalsofthegroup.
offeelingsoffusion. Transientcontextualoriginsoffusion. Onecommoncause
Fusion-friendly ideologies. One tradition that might rein- of local fusion is sharing “bonding” experiences with others in a
force or amplify biological precursors of fusion is the Jus- group context. This may encourage the belief that fellow group
Sanguinis citizenship tradition, wherein the main criterion for membersperceivetheworldinafundamentallysimilarway(e.g.,
membershiptothenational“polity”isbloodline.Thistraditionis Pinel, Long, & Crimin, 2008, 2010; Pinel, Long, Landau, Alex-
relatively common in countries such as Germany and Austria. ander, & Pyszczynski, 2006), a perception that may produce the
Lowerfusionrateswithcountryshouldemergeincountrieswitha profound feelings of connection that give rise to fusion. There is
Jus Soli citizenship tradition, where bloodline criteria are irrele- also a relatively “hot” pathway to local fusion in which group
vantwhendeterminingcitizenship(e.g.,France;seeKohn,1944; members share significant experiences with others and subse-
A.D.Smith,2001). quently fuse with them. For example, the American soldier in
Although bloodline considerations are less prominent in the
Afghanistan quoted at the outset of the article fused with his
ethic Ubuntu, it likewise encourages fusion-like feelings of alle-
comrades in arms after co-experiencing a series of frightening
giance toward other group members. Prominent in the Bantu
situationswiththem.Inthisinstance,thesoldiercametoperceive
dialectinSouthAfricaandneighboringcounties,Ubuntuempha-
hisfellowgroupmembersaskin-like,referringtothemas“broth-
sizes the fundamental connectedness of group members and is
ers,”despitetheabsenceofabiologicalrelationshipwiththem.In
considered the essence of being human. It refers to a philosophy
fact, some evidence suggests that sharing subjective experiences
that encourages compassion and generosity toward fellow group
with others may be a more powerful predictor of attachment to
membersaswellaseffortstoachievemutualunderstanding(Tutu,
fellow group members than the perception of shared objective
1999). If such efforts fail and misunderstanding emerges, recon-
qualities (Atran, 2010; Drury, 2011; Pinel et al., 2008, 2010,
ciliationispreferredoverretribution.
2006). Moreover, shared experiences seem to be particularly po-
Fusion-like allegiances with country are also apparent in the
tent in facilitating attachment to others when the experiences are
societiesinwhichthereisastrong“cultureofhonor,”suchasin
challenging or traumatic rather than positive (Baumeister, Brat-
certainMediterraneancountries(RodriguezMosquera,Manstead,
slavsky,Finkenauer,&Vohs,2001).
&Fischer,2000,2002)andinpartsofthesouthernUnitedStates
Considerparticipationinritualssuchastheordealsofinitiation
(Cohen & Nisbett, 1994; Cohen, Nisbett, Bowdle, & Schwarz,
cults,millenariansects,andvisionquests.Such“imagistic”rituals
1996). Such societies emphasize a strong linkage between the
(Atkinson & Whitehouse, 2011; Whitehouse, 1995, 2000, 2004)
honor of the individual self and the honor of fellow ingroup
aretypicallyemotionallyintenseeventsthatareexperiencedrarely
members. The relational ties between ingroup members are so
(only once in a lifetime in some cases). The intensity of such
important in such cultures that it is normative to respond to
rituals is exaggerated by extreme forms of deprivation, bodily
challengestothehonorofingroupmemberswithextremeretribu-
mutilation and flagellation, and psychological trauma based
tion,includingmurderingtheoffendingparty.
around participation in shocking acts. These practices are wide-
In addition to encouraging the development of relational ties,
spreadinsmall-scaletribalsocieties(Whitehouse,1996),modern
fusion-friendlyideologicalsystemsmayalsoexplicitlyencourage
rebelgroups(Whitehouse&McQuinn,inpress),andsomeancient
groupmemberstolabeltheirtiestooneanotherasrelational.For
civilizations (Whitehouse & Hodder, 2010). Experiments show
example,militaryunitsthroughouttheworld,fraternities/sororities
thatimagisticritualstypicallyinvolveintrinsicallypuzzling(caus-
in the United States, and businesses in Japan all advertise their
ally opaque) procedures that trigger intense reflection (Richert,
groups as being family like. Such labeling practices encourage
Whitehouse, & Stewart, 2005). Such reflection appears to be an
groupmemberstoconstruethegroupascloselyknitandfusewith
it.Similarly,insomeculturesandcountries,thedominantideology essential element in the process (cognitive dissonance reduction
emphasizesthatthecountryislikeafamilytoitspeople.Suchwas doesnotappeartoadequatelyexplaintheseeffects),6asreflection
the case in former communist countries like the former Soviet producesenduringandvividepisodicmemoriesfortheordealsand
Union,whereinfellowcitizensreferredtoeachotheras“brothers” theothergroupmemberswhouniquelysharedintheritual.Tothe
and“sisters.”Evenmoreexplicitly,inChina,Confucianideology extentthathumansare“creaturesofconcreteness”(Nisbett,Bor-
(reinforced by Chinese leaders) emphasized that the nation is gida, Crandall, & Reed, 1976), such experiences may be particu-
“familywritlarge”(Liu,Li,&Yue,2010).Itisthusnotsurprising
thatrecentevidencehassuggestedthatfusionratesarestrikingly
6Peopletheoreticallyexperiencedissonanceonlyinsofarastherelevant
highinChina(75%amonguniversitystudentsaccordingtoJetten,
actionsarefreelychosen(e.g.,Aronson&Mills,1959;Gerard&Mathew-
Go´mez,Buhrmester,Brooks,&Swann,2012).Incontrast,fusion
son,1966).Assuch,dissonancetheoryispoorlysuitedforexplainingthe
rates were much lower in Western countries such as Australia
tendencyfortheseparticularpainfulgroupinitiationstoproducecohesion
(10%) and the United States (25%). Countries that emphasize becausenovicesundergoinginitiationintobellicosetribesareoftenforc-
relationaltiesbutlackastrongandsharednationalideologyfellin iblyabductedbytheirinitiatorsandareobligedtosubmittotheritualson
between(60%inIndonesia,32%inSpain). painofdeath(Barth,1987).
450 SWANN,JETTEN,GO´MEZ,WHITEHOUSE,ANDBASTIAN
larly compelling sources of fusion—sometimes resulting in alle- another group member in a manner that is unforgivable (e.g.,
giancesthatarestrongerthanthosefoundbetweenbloodrelatives. causing a group member’s death), ejection from the group and
Indeed, Atran and Henrich (2010) have suggested that group de-fusionmayfollow.Finally,peoplemaychoosetode-fusefrom
rituals encourage people to make and keep oaths to each other the group if they conclude that it has changed in a way that
(Boyer,2001)andtosuppressselfishnessandfree-riding(Noren- contradictsitscorevaluesandbeliefs.Forexample,someAngli-
zayan & Shariff, 2008). In this way, rituals may give rise to a canpriestscametobelievethatthechurch’sdecisiontopermitthe
psychological immune system (Wilson, Gilbert, & Centerbar, ordinationofwomenwouldchangethenatureandidentityofthe
2003)thatprotectspro-groupbeliefsagainstrationalcounterargu- Church. They accordingly left the Church because they believed
ments. thatitsidentityhadchangedinamannerthatwassofundamental
Whatever the precise nature of the mechanism underlying the thatitceasedtobethechurchthattheyhadjoinedoriginally(Sani
effects of rituals may be, this discussion of variables that may &Reicher,1999).
promotefusionleadsonetoaskiftheremaybeacorrespondingset
of variables that undermine fusion, a process we have dubbed
“de-fusion.” In the section that follows, we consider when such Discussion
“de-fusion”processesmayoccur.
This article focuses on “identity fusion,” a form of alignment
withgroupsthathaspreviouslybeenoverlooked.Unlikeconven-
Causes of De-Fusion
tional forms of alignment with groups, fusion is marked by a
Theirrevocabilityprinciplesuggeststhatoncepeoplefusewith visceralfeelingofonenesswiththegroup.Thisfeelingisassoci-
a group, they will tend to remain fused with that group. Indeed, ated with a highly agentic-personal-self and unusually porous,
instances of de-fusion are likely to be emotionally wrenching, as highlypermeablebordersbetweenthepersonalandsocialself.The
they theoretically entail substantial restructuring of the self- porousbordersassociatedwithfusionraisethepossibilitythatboth
concept, one’s relation to others, and even the very meaning of the personal and social self will combine synergistically to moti-
one’s actions (Wegner & Vallacher, 1986). For these reasons, vate unusually extreme sacrifices for the group. In addition, the
peoplemaydefusefromagroupquitereluctantlyandonlyifthey porous borders will encourage highly fused persons to develop
feelthatremaininginthegroupissimplyimpossible. strong relational ties with their fellow group members and main-
Onepotentialcauseofde-fusionisthedisbandingofthegroup tainlastingcommitmenttothegroup.
or completing one’s terms of service with the group. On such Fusion theory specifies four principles: agentic-personal-self,
occasions, highly fused individuals will be strongly motivated to identity synergy, relational ties, and irrevocability. The research
restore their relation to the group. Consider, for example, the literature has provided support for each of these principles. Al-
protagonist in the film The Hurt Locker. As the film opens,
thoughmostofthesupportforfusiontheoryfocusedonthedegree
viewers are taken to Iraq where they are acquainted with an
towhichfusedpersonsendorsedextremebehaviors(i.e.,fighting
American demolition expert who becomes fused with his com-
anddyingforthegroup),fusionwasalsolinkedtoovertbehaviors
rades in arms as he repeatedly risks his life in the service of his
suchasdonationsofpersonalfunds(Go´mez,Morales,etal.,2011;
country.Whenhistourofdutyendsandheisobligatedtoreturn
Swann, Go´mez, Dovidio, et al., 2010), motor behavior (Swann,
home,heisreunitedwithhisfamily.Yet,heisuneasy,forhelongs
Go´mez,Dovidio,etal.,2010),andactualalterationofprimarysex
for the world he left behind, especially the men with whom he
organs within a sample of transsexuals (Swann et al., 2012).
fought. In the concluding scene, he volunteers for another life-
Moreover,fusedparticipantsendorsedplungingthemselvesinthe
threatening tour of duty in Iraq, apparently because he could not
path of a runaway trolley to save the lives of individual group
defusewiththemilitary.
members(Swann,Go´mez,Dovidio,etal.,2010).
Although The Hurt Locker is a fictional account, there can be
Giventheutilityoffusioninpredictingpro-groupbehavior,itis
littledoubtthatpeopletypicallyfinddefusingfromagrouptobe
important to learn more about its origins. Evolutionary theory’s
psychologically wrenching. Such instances point to a liability of
principle of inclusive fitness might explain the self-sacrificial
fusion.Thatis,becausethestateoffusionissoallconsuming,it
behaviorsofhighlyfusedpersonsifoneassumesthathumansrely
may compromise people’s capacity to compartmentalize their
on tribe-detection systems that are prone to false positives in the
group-related experiences (Hugenberg & Bodenhausen, 2004).
modern world. At a more proximal level, some cultures and
The single-mindedness of highly fused individuals may thus im-
ideologies promote beliefs, attitudes, and norms that motivate
pair their ability to display sufficient role flexibility needed to
maintainhealthyrelationshipswithindividualswhoarenotmem- people to fuse with the group and its members. Fusion and self-
bers of the fused group. In addition, insofar as such individuals sacrifice may also be triggered by transient contextual variables,
display extreme pro-group behaviors that are deviant in nature, such as certain ritual practices. Tests of these hypotheses will
they may find that they are not welcome in circles outside the clarify how biological, cultural, and contextual factors interact to
fused group. For example, gang members who commit crimes as produce,sustain,andchangefusion.Insodoing,researcherswill
partofinitiationritualsmaybecomesocialpariahswhocannotwin notonlyattainaclearerconceptionofthenatureoffusionandhow
acceptancewithinalternativeopportunitystructuresthatwelcomed itmotivatesextremebehavior,theywillalsolearnmoreregarding
thempriortotheirfusionwiththegang.Thismayfurtherreduce howfusionissimilarto,anddifferentfrom,otherformsofalign-
thelikelihoodthatindividualswilldefusefromthegroup. ment with groups such as identification. Such knowledge will
De-fusionmayalsooccurwhenrelationaltieswithgroupmem- advance understanding of each of these constructs as well as the
bers are shattered. For example, if one group member betrays processesthattheyregulate.