You are on page 1of 16

Chapter3

An lntegrative
Theoryof
lntergroupConftict
Henti Taifet
John Tumer
Universiuof Btistol,England

INTRODUCTION D. T. Campbetl( 1965)asthe ..realistic group conflict


theory" (R.C.T.). Its pointof deparrurelortireexpla_
The aim of this chapteris to presentan outline of nation of intergroupbehavioris in what Sherif(1966)
a theory of intergroup conflict and somepreliminary has called the functional relations between social
data relating to the theory. Fint, however, this ap. groups. Ib cenbal hypothesis_..real conllict
proachto intergroup behavior and intergroup confli-ct of
gT.op-^T:oo! !*ses intergroup conflict" (camp.
must be set in context, in relation to other approaches beII,.1965,p.287)-is
deceptivetysinple, intuitiveiy
to the sameproblems. convin_cing, andhasreceivedsrong empiricatsupport
Much of the work on the social psychology of (including Sherif & Sherif, 1953; Avigdor, i{i53;
rntergrouprelations has focused on patternsof indi_ H:rvey, 1956; Sherif, Harvey, $hlte, Hooa, A
vidualprejudiceanddisqiminationandon themotiva- Sherif, I961; Blake& Mouton,I96la, 1962c;Bass&
tional sequenc€sof interp€rsonalinteraction.Out_ Dunternan,1963;D. W. Johnson,1967;Diab,1970;.
standingexamplesof theseapproaches can be found, R.C.T. waspioneeredin socialpsychologyiy
respectively,in the theoryof authoritarianpersonality _ -
the Sherifs, who provided both an ctiotogy ofitrd
(Adorno, Frenkel-Brunswik,Levinson,& Sanfor4 grorp hostility anda theoryof competitionasrealistic
1950)andin the various versionsandmodificationsof and inst-umentalin charact€r,motivatedby rewards
the th:ory of frr'stration,aggression,and displace_ which,
in principle, are extriNic to the intergroup
ment (such as Be*owitz, 1962, L969, fgZli. ttr"
situation (see.Deufsch, l949bi Julian, l96gyl Op
commondenominatorof most of this work has been posed
group int€resb in obtaining scarce r€sour@s
the stess on the intraindividual or interpersonal promote
psychological processes leading to prejudiced _competition,and positively intexd€pendent
at_ (superordinate)goatsfacilitate cooperation.Conflia_
utudes or discriminatory behavior. The comolex
ing_interesbdevelop through competition into overt
a social conJlict. It appears,too, that intergroup com-
wls_s9_99@_sog!4_processesEEGEEii petition cnhancesintagroup morale, colesivcness,
3.-g-[:A*dth"="1.""h"td;ffi ; 3:d-cooperatiotr(seeVinacke, 1964;Fiedler, 1967;
ur rne rocus ol the social psycholoqisb,
Dreoccu- Kalin & Marlowe, 1968). Thus, rhe real conflicts of
group int€r€stsnot only ceate anogonistic intergroup
rclations but also height€n identification *ith, uni
The alternative to these approachesis rep- positiveattachment
rcsentedin the work of Muzafer Sherif and to, thein-group.
his as_ This identrfication with 6e in-group, however,
sociates(seeChapter 1), and has beenrcferred
to bv has been givcn relatively little promiDence.in the
34 Chap@r3

R.C.T- asa lheoreticalproblemin its own right. The extremeareprovidedby the behaviorof soldiers from
developmentof in-group identificationsis seenin the opposingarmiesduring abatde,or by thebehavior ata
R.C.T. almost as an epiphenomenon of intergroup negotiatingtable of membersrepresentingtwo-parties
conflict. As reated by the R.C.T., theseidentifica- in an inlenseintergroupconflict.
tiots arcassociatedwith certainpattemsof intergroup The main empirical questionsconc€rnthe con-
relations,but the theorydoesnot focus either uponthe ditions that determinethe adoptionof forms of social
proc€ssesunderlying the development 8nd mainte. behavior nearing one or the olher exteme. The
nancc of group identity nor upon the possibly au- first-and obvious-answer concernsiltergroup con-
tonomous effects uPon the in-group and intergroup flict. It canbe assumed,in accordance with our com-
"subjective" aspects of group mon experience,that the moreintenseis an intergroup
behavior of these
membenhip. It is our contentionthat the relative ne- conflict, the morc likely it is that the individuals who
glect of theseprocesses in the R.C.T. is responsible arc membersof the opposinggroupswill behavetG
for some inconsistenciesbetween the empirical data ward eadrother asa function oftheir respectivegroup
"classical " form. In this sense, memberships,rather than in terms of their individual
and the theory in its
the theoretical orientation to be oudined here is in- characteristicsor interindividual relatiorships. This
tendedtrot to replac€the R.C.T., but to suPPlement waspre.isely why Sherif (1966,for example)was able
it in some rcspects that seem to us essential for an to abolish so easily the interindividual friendships
adequatesocial psychology of intergroup conflict formed in the preliminary stagesof some of his field
-partiqrlarly as the undentanding of the Psycho- studieswhen, subsequently,the individuals who had
logical aspectsof social changecannot be achieved becomefriendswereassigned to opposinggroups.
without an appropriateanalysisof the social psychol- An institutionalizcd or explicit conflict of "ob-
ogy of social conflict. jective" interesb betweengroups,however, does not
provide a fully adequatebasis, either theoretically or
THE SOCIALCONTEXTOF INTERGROUP cmpirically, to accountfor many si ations in which
BEHAVIOR the socialbehaviorof individualsbelongingto distitrct
Our point of deparhre for the discussion to grcups can be observedto appmach the "group"
follow will be an a priori distinction between two exteme of our continuum. The conflict in Sherif's
ext€mes of social betravior, correspondingto what studies was ' 'institutiondized.' ' in that it was offi
we shall call llllerpersonal versus intergrouP be' cially arrangedby theholiday campauthorities;it was
"explicit" in that it dominatedthe life ofthe groups;
havror.At oneexueme(whlchmostproDaolycanDot
"real life") is the and it was "objective" in the sensetlat, by the terrns
be found in is "pure" form in
hteraction between two or more individuals that is of the competition, one of the groups hal to be the
lJly determinedby their interpersonalrelationships winner andthe othergroup the loser. And yet, thercis
and individual characteristics.and not at all affected evidencefrom Shcrif's own sfirdies atrd fiom other
by various social groups or categoriesto wbich they researcl (to which we shall refrrn lat€r) that the in-
respectively bclong. The other extreme consists of stitutionalization, explicitness, and "objectivity" of
interactions between two or more individuals (or an intergroupconflict arenot r€ceJ.raryconditions for
groups of individuals) which are/zlly detcrminedby behavior in terrns of the "group" exteme, although
their respeclivemembershipsin varioussocial groups drey will often prove to be sufrcient conditions. One
or categorics,and not at all affect€d by the interin- clear exanple is provided by our carli€r experimenb
dividual personal rclationships betrwccnthe people (fajfel, 19?0; Tajfel, Billig, Bundy, & Flament,
involvcd. Hereagain, it is unlikely that "pure" forms 1971),which we shall discussbriefly below, in which
of this cxtrcmc can be found in "rEd" social situa- it was found that intergmup discrimination existedin
tions. Examples nearing the interp€rsonal cxtrcme conditionsof minimal in-group affiliation , anonymity
would be the relations betwecn wife and husbandor ofgroup membership,absenceofconflicts of inter€st,
between old fricnds. Examples near the intergroup and absenceof previoushostility betwcenthe groups.
An lntegrative Th@ryof lntergrcup hnflict
Other social and behavioral continua are ployedduring the depressionof the 1930s)that
associat{ with the interpersonal_intergroup con- the
impossibitityof ..geaingout" on one's own, as arl
tinuum. One of them may be coruidered as havins individual, becomesan everydayreality thdt det€r_
a causal function in relation to the interpersonal_I mrnesmany forms of intergroup social bdravior- But
intergroup continuum. We shall characterizeit asain even this exampleis still relatively extreme. Many
by its two extremes,to whidt we shall refer as 4o- socialintergroupsituationsthatcontain,for whatever
cial mobility" and "$ocial change." These teifi reasons, stong elementsof stratificati on perceivedas
are Dor usedherc rn theL sociologicalsense.They srct will move social behavioraway from the pole of
refer instead to the individuals, belief systems interpenolat patternstoward the pole of intergroup
about the nature and the structure of the relations
between social groups in their society. The belief Pure*:.@'
systemof "social rnobitity" is basedon the general
assumptionthat the society in *,trich the individuals this belief system is that, in tte relevant intergroup
live is a flexible and permeableone, so that if thev are situatiotrs,individualswill not int€ractas individuals.
Dotsatisfied,for whateverreason, with thecondiions on thebasisof theirindividualcharacteristics
imposed upon their lives by membership in social or inteF
personalrclatiorships, but asmernbersof theL groups
groulx or social cat€goriesto which they belong, it is standingin c€rtain defined relationships to memb€rs
possible for them @e it through talent, hard work, of othergmups.
good luck, or what€ver otrer means) to move indi_ Obviously, one must exp@t a marked correla_
vidually into anothergroup which suits them beuer. tion betweenthe degreeofobjective sEatification in a
A good example of this system of beliefs, built into social system(howevermeasurcd)and the social dif_
the explicit cultural and ideological traditions of a fusion and intensity of the belief system of ..social
society, is provided in the fol.towing passagefrom change." This, however, cannotbe a one-to_onerela_
Hirsclman(1970):
tionship for a numberof reasons, some of which will
be discussedbelow, although we cannotin this cha}.
lbc EaditionalAmericanide-aof successconfirrnsthc hold
.whidl cxit hashad on the D8tional ter go i nlo thedetailsof themany social_psychological
iergination. Succcss conditionsthat may deGrminette bansition in crrtain
----q, what amountsto thc same thing upward
, social social groups from an acceptanceof stratification to
Eobility-has lotrgbccncotrccivcdin terEsof evolutionary
individualism.Thcsuccessfulindividua! behavior characteristicof the int€rgoup pole of our
\to starrsoutat a
low rung of the socialladder,neccssadlyleaveshis own first conti.nuum-that is, to lhe sreation of social
groupasheriscs;he ..passes " into,or is ..acrepted,,
by, movemetrtsaiming to change (or to preserve) the
lhcncxthighergoup. Hc tak6 his immcdiatefamily alon!, status quo (tajfel, in press). It may be interesting,
buthardlyanyonc elsetpp.lO8-1091. however, to point to the close relationship that exislts
betw€enan cxplicit intergroup conflict of interesC on
At the other extrcme, the bclief systemof . !o- ,
the one hand, and the ..social change" system of
cial change" impli€s that lhe naturc and sfiuctur€of beliefs on the other. One of the rnain featuresof this
the relations betwe€o social groups in the society is belief systemis the perceptionby the individuals con-
p€rcciv€d as chrracterized by markcd statificad;n, cemed that it is iurpossibleor extremely difficutt to
naking it impossibleor vcry difficxlt for individuals, move individually firom thcir own group to another
as individuals, to invest themsclvcsof an unsatisfac_ group. This h precisely thc situation in an intense
tory, underprivilcged, or stigmatizcd group membcr- intergroup conflict of intaests, in which it is cx-
ship. A castc system, bascd on tace or anv other ftmely diftiodt for an individual to conc€ive of the
criteria perccivcd as immutablc, is an obvious cxam_ possibility of "beraying', his or her own group by
ple, but it nced not be thc main cxample. The moving to thc opposing group. Although this does
economicor social r€alities of a society may be such happenon occasion,sanctionsfor sucha mov€are, otr
(as, for example, in the caseof the millions of unem- the whole, powerful, andthc value systems(at leastin
Chaptet3

our cultures)are in flagrant oppositionto it. To usean oulgroup as undifferentiateditems in a unified social
exampleftom social-psychologicalresearch,it seems ""E;,r,ffir
hardlypossiblethat oneof theboys in Sherif's holiday charact€ristics. The vastliteraturein socialpsychbl-
campswould decideto changesides, even though oli-ofr-fi6Eictioning of group stereotypesin situa-
some of his previously contractedfriendships over- tions ofintense intergrouptensionsis no more than an
lappedgroupboundaries. exampleof this generalstatement.
The intensity of explicit intergroup conflicts of Thus, this preliminary conc€ptualization rep-
interestsis closely related in our cultures to the de- resentsan approachto the social psychologyof inter-
gree of opprobrium attachedto the notion of "rene- group rclations that takesinto accountsoci4lE)alities
gade" or "traitor. " This is why rhe belief systems aswellastheirreflection-li6iiiiTdiifiGiJh
conespondingto the "social change" extremeof our .
continuumare associatedwith intenseintergroupcon-
flicts. Theseconflicts can be conceived,therefore, as just discussed:at the beginning,becauseit can be
creatinga subclassor a subcategoryof the subjective assumed without much difficulty that the "social
intergroup dichotomization characteristicof that ex- change" belief system is likely to reflect either an
treme of the belief continuum. They sharethe basic existing and markedsocial stratification or an intense
feanrre of the "social change" system of beliefs, in intergroup conflici of interests,or both; at the end,
the sensethat the multigoup sFuctureis perceivedss becausethe consequencesof the systemsof beliefs
claracterizedby the eKteme difficulty or impossibil- arising from fte social situatioDsjust mentioned are
ity of an individual's moving from one group to likely to appearin t}le form of unifred group actions
another. -lhat is, in ttre form of social movementsairning
Tbe continuum of systemsof beliefs disorssed either to createsocial changeor to prcs€rvethe status
so far has be€n seen to have a causal function in quo. We shall return later to an elaboration of the
sbifting social behavior toward menbers of out- kinds of hypothesesthat can be put forward concern-
groupsbetweenthepolesof''iDterpersonal" and ''in- ing the creation of change versus the preservation
tergroup" behavior. To concludethis part of our pre- of status quo. But before this is done, the realistic
liminary discussion, we mustcharacterizebriefly two group conflict theory must be consideredagainstthis
firther and overlapping continua, which can be con- generalbackground.
sidercd as encompassingthe majot consequencesof The implications of this conceparalizationfor
social behaviorthat approaclesoneor the otherend of intergroup relations in stratifi€d societies ard insti.
the interpersonal-int€rgroupcontinuum. They both tutions are both evident and direct. Whenever so-
have to do with the variability or uniformity within a cial stratification is basedupon an unequal division
group of behavior and attitudes concerning the rele- of scarce resources-such 8s power, prestige, or
vant out-groups. The first may be describedas fol- wealth-betwe€n social groups, the social situation
lows: the nearcrare membersof a grouo to the ' 'social should be characterizedby pervasive ethnocentrism
change'' extremeof the belief-systernscotrtinuum and and out-grcup antagonismbetweenthe over- and un-
the int€rgroupexFemeof the behavioralcontinuum. derprivilegedgroups(Oberschall,1973, p. 33; see
ggfe.nprmity they will show in their behavior also Chapt€rs 5 and f2). Howe-v-er,decadesof rc-
E
iowardmemb@- search into cthnic-group relations strggestthat eth-
funcse nocentsismamongsratifi eCgroupsis]6iii-i6i]ilras
continua will be correspondingly associarcd with b€€n, v€ry much a one-waystreet.Two rcc€ni book
great€rin-group variability of behavior toward mem- (MGAJy/5;cil-6Fftives6'd, 1976)surnrnarize
bcrsof the out-group.The seaondstaGmentis closely a great deal of cvidencethat minority or subordinate
r€latedlo lhe first the nearerarc merrbirs of a group group.members-such as lbe American Blacks, the
to the'social chatr=e-Fdl5d--lnterEmup--€x- French Canadians,the New Zealand Maoris, or the
e South African Bantus-frequently tend to derogate
An lntegrative Theoryot lntergroup Conflict
:ial the in-group and display positive attitudestoward the
ual the-processesof change bccome very rapid.
dominantout-group. In other words, deprivedgrcups 11hen
Stanrsdifferenc€sbetweengroupsoflen do nitreiain
ol- 8re not always elhnocentric in the simple meaningof unilaterallyassociatedwith low levelsof intergroup
ua- the tccm; |lley may, in fact, be positively oriented conflicb. For example, the generalization
an toward the depriving out-group. Data of this kind are madl
above-that certainformsof political,economic.
notconsistent with asimpleapplicationofthe R.C.T. and
social subordinationof a socialgroup tend to eiimi_
cp- Sornewriters (including Gregor & Mcpherson,
]lt-
nate or even reverse its ethnocentrism_is already
1966;Modand, 1969;Milner, 1975,p. 93) havear_ dated. Researchconducted in the last decade
ies gued that the status relations betwe€ndominant and or so
rcveals a changing pattern in intergroup relations.
rgh subordinate groups deterrnine the latters' identitv American Blacks (see N. Friedman, 1969; Hraba
fs. problems. (By sociat stlo,tuswe mean a .anking oi &
Grant, l9?0; Paige, 1970; Brigham, f97l; Harris
lc€ hierarchyof perceived prestige.) Subordinategroups & Braun, t97I), Rench Carudians(Berry, Kalin, &
b€ often se€mto intemalize a wider social evaluation of Taylor, 1976), New Zealand Maoris
:id lhemselvesas "inferior" or ..secondclass," and {aughan, in
an this consensual inferiority is rcproduced as rela_ ry-To,_utd rhe Wetsh (Bourhis, Giles, & Tajfel,
1973;Giles & Powesland,1976;,for instancc,now
nsc tivc self-d€rogationon a numberof indices that have se€m to b€ lejecting their previously negative in_
od, been uscd in the various studies. Consensualshtus group evaluations and developing a positive e6no-
cfs iself-where subjeclive and accorded prestige are c€ntrtc group identity. This .construclion of positive
AI€ identical-is problematic for the R.C.T., which con_ in-group attitudes is often accompaniedby a
)ns ceptualizesprestige as a scarccresource,like wealrh new
militancy over political and economicobje,ctives(see
ing or F)wer. Status differences betwean groups, like Tomlinson,1970).
.nlS othcr inequalities,should tend to acc€trtuat€ lhe inter- But &ese developments do not rescue the
lhc group conflics of inter€s$. Therefore, acaording to
R.C.T. in its original form. The very suddennesswith
m- tbe R.C.T., the impact of low statusupon a suboidi-
ion nate group should be to illensify its antagonismto- t
;tic objective deprivation and tl'e e
wad tfie high-status group (fhibaut, teSOy. yet,
his undersomeconditionsat lest, low socialstatusse€rns @ul
to be conelated with an cnhancement,rather than a 9rd@99s4cEs. on-E" "*Gry,
for uerc rs otten less ..objective" deprivation than there
lessening, of positiveou&groupaftirudes. has been in the past. Atr active and new searchfor a
tti- It could be arguedthat only conllicrs of inter€st
sG.
positive group identity seems!o be one of the critical
p€rc€ivedas such create hostility. This requhes
&at lactors r€sponsible for the reawakeaing of these
ion groupsmustcomparetheir respectivesitrations . And
, groups' clains to scarc€resourc€s(Dizad, 1970).
o( accordin€to some views, it is only dativcly similar
ion gmupslhat engagein mutual comparisons;thcreforc, _ . In summaqr, the R.C.T. states that opposing
sm
clains to scarc€rcsourcrs,suchaspower, prortig", oi
many forns of stau$ diffcrences will reduce per_ wealth, generateethnoccntrismand ant gonis; be-
tn- c€ivcd similadry (sce Fe tinger, f 954; Kidder
tce
e twc€n grcups. Thcrefore, low stafis should tend to
Jtcwart, 1975). It follows that sanrs systems rnay
rc- intensiry out-goup hostility in groups that 8re politi_
reducesocial conflict by resticting thc rangeof
lh-
mean_ caxy, cconomic€lly, or socially subordinate. The
colnoisgns availablc ro any given goup. This evloenc€ suggesb, howwer, that whcre social-
ns lnSt{
qypottsls may be I uscfi.rl
lool aosccormtfor soc of shuctural differcnc€s in the distribution of rcsources
,ks thc-deErminatrbofsocial stability; but if it is taken
nc to havc b€eninstibtionalized, lcgitimizrd, andjustifried
Its logical conclusion, it can ac€ountfor no morc than
atc tbrough a consemually acc€ptcdst8t|s system (or at
that. It fails to accountfor social changc(in the
the
scnse reasta staorsEystemtrat is sufficiently frrm andperva-
of changes in the mutual relations, bchavior,
the and sive to preventthe creationofcognitive altemadvesto
attitudesof large-scalehuman groups that have
ate
be€n it), the resulthasbeenlessandnot moreethnocentrism
dlstlnctly different in status in the past), panicularly
in the different statusgroup6. The price of this has
A lnEgrctlve Theoryot tntergroup @nflict
:i8l &e in-group and display positive auitudestoward the when the-processesof changebecomevery rapid.
ual dominantout-gmup. In otrer words, deprivedgroups Sta[rs differcnc€sbetwe€ngroupsoften do notrentain
ol- are not always ethnocenfic in the simple meaningof unilatcrally associatedwitr low levels of intergroup
ua- fie tcrm; ftey may, in fact, bc positively oricnted conflicb. For example, the generalizationmadl
an toward the depriving out-group, Daraof this kind are above-lhat c€rtainforms of political, economic,
notconsistent wift a simpleapplicationoftheR.C.T. and
social subordinationof a socialgroup tend to eiimi_
q' Some$Titers (including Gregor & Mcpherson, nate or even reverse its ethnoc€ntism_is already
Dt- 1966;Modand, 1969;Milner, 1975,p. 93) havear_ dated. Researchconducted in the last decade or
ies gued that the status relations betweendominant and sl
reveals a changing pattern in intergrcup relations.
lgh subordinale groups determine the latt€rs' identitv American Blacks (see N. Friedman, 1969; Ilraba &
fs. problems. (By sociat Jbt[.r we mean a ranking or Grant, 1970; Paigc, 1970; Brigham, l97l; Harris
lc€ hierarchyof perceived prestige.) Subordinategroups & Braun, t97I), Rench C:nadians(Berry, Kalia, &
b€ often s€cmto intcrnalize a wider social evaluation of Taylor, 1976), New Zealand Maoris (Vaughan, in
:id drcmselves as "inferior" or ..second class," and
ln this consensual iofcriority is rcproduced as rela_ ry-.),_gd the Wetsh @ourhis, ciles, & Tajfet,
1973;Giles & Powesland,1926),for instance.now
nsc tivc self-derogationon a number of indice that have se€m to be rejecting their previously negative in_
0d, been used in the various studies. Consensualshtus group evaluations and developing a positive ethno_
cfs i8ef-whcre subjective and accorded prestigc arc c€ntnc group identity. This construction of positive
are identical-is problematic for the R.C.T., which con_ itr-group attitudes is often accompaniedby a new
)ns ceptrslizes prcstige as a scarceresourc€,like wealth militaacy over potitical and economicobje,ctives(see
ing or power. Stanls differenc€s between groups, like Tomlinson,1970).
Ars other inequalities, should tend to accentuatethe int€r- But these developmenE do not rescue the
lhe group conflics of hter€sb. Therefore, according to
R.C.T. in its originat form. The vcry suddennesswith
ln- the R.C.T., thc irnFact oflow statusupon a subordi_
ion nate group should be to int€nsify its antagonismto_ t
rtic oljdve deprivation and rherdor ne
ward thc high-sh0rs group (fhibaut, 1950). yet,
his undersomcconditions at least,low socialstat$ seems @uI
to be correlated with an enhanc€ment,rather than a
gu$@eeesun.
..objective"
on-E" *oE y,
for rDerc$ ottcn less deprivation than there
lessening, of positiveout-groupattitudes. has been in the past. Atr aclive and new searchfor a
rti- It could be arguedthat only conflicts of interest positive group identity scemsto bc one of the critical
to- p€rc€iv€das such createhosdfiry. This requires that
tactors responsible for the reawakaning of these
ion grcup'smustcomparetheir respectivesiO.rations . And , goups ' claims to scarce!€sourc€s(Diz:d I 970).
q
accordingto some views, it is only telatively similar ,
ion groupsthat engagein mutual comparisons;thcrefore, , _ In summary, the R.C.T. states that opposing
clair.nsto scarcercsourc€s, suchaspow€r, prcstige, oi
sm many forrrs of staus diff€rences will rcduce oer_
ln- wealth, gcnerateethnocentrismand anhgonisi be-
ceivcd similariry (see Fcstinger, l95a; Kiddei & twe€n gmups. Thcrefore, low stat$ should tend to
tog stewart, 195). It follows that statussystemsrDay
te-
intensify ourgroup hostility in grroupsthat are politi_
reduccsocial conflict by restiixing the rangcof urean_ cat-ty, cconomically, or socially subordinate. The
rh- ugur-comparisonsavailableto any givcn group. This cvidence sr,rggess, howcver, that where social-
ras lypot[csa may be I useftl tool to accauntfor som of stuctrral diff€rcnces in thc distribution of resources
*s &e_dcbrminants of social stability; but if it is takcn !o
'tzc have bccn instiurtionalized,lcgitimizcd, andjust'rfied
tB logical conclusion, it can acaountfor no morc than
atc ulrougn a @nsensuallyacceptrd 6tahs system (or at
thrt. It fails to accountfor social change(in the sense
tle leasta staarsB)6temthat is sufficicntly firm andp€rva_
of changcs in the mutual relations, bchavior,
the and sive to pr€ventthe creationofcognitive altcrnativesto
aftihdes of large-scalehuman groups that havc be€n it), the resulthasbern lessandnot moreethnocentrism
a@ dlstindly different in shrus in the past), panicularly
in the differcnt statusgroups. The price of this has
Chapter3

been the subordinategroups' self-€ste€m.On the evaluationsand behavior,Not only are incompatible
other hand, whenevera subordinategroup begins, for group intefesb not alwayssufficient to generatecon-
whatever reasons,to question or deny its presumed flict (as concludedin th€last section), but there is a
characteristicsassociatedwith its low status, this good deal of expetimentalovidencathat thesecondi-
seemsto facilitate the reawakeningof a previously tions are not alwaysnecessaryfor the developmentof
dormant conflict over objective resourc€s. At the competition and disqimination betweengroups (for
sametime, it is likely that oneof the counterrcactions example,Ferguson& Kelley, 1964;Rabbie& Wil-
from the dominantgroupsin suchsituationswill be to kens, l97l; Doise & Sinclair,1973:Dojse & Wein-
work for the preservationof the previously existing berger, 1973).This doesnot mean,of course, that
"subjective" and "objective" differertiations. in-groupbias is not influencedby the goal relations
A tentativehypothesisaboutintergroupconflict betweenfre groups(seeHarvey,1956).
in sratified societiescan now be offered:An unequal All this evidenceimplies that in-group bias is a
distribution of objective resourcespromotes antag- remarkably omnipresentfeatrre of intergroup rela-
onism between dominant and subordinate groups, tions. The phenomenonin iis extremeform has been
provided that the latter group rejecb its prcviously investigatedby Tajfel and his associates.There have
acc€ptedand consensuallynegative self-image, and be€na numberof studies(Tajfelet al., l97l; Biuig &
with it the statusquo, and starts working toward the Tajfel, 19?3;Tajfel & Billig, 1974;Doisc, Csepeli,
developmentof a positive group ideotity. The domi- Dann,Gouge,Lanen, & Ostell,1972;Turner, 1975),
nent group may reactto thesedevelopmenbeither by all showing that thc mere perceptionof belonging
doing everything possibleto maintain andjustify the !o two distinct grouDs-rhat is, so-Al-eaEEbfiZatbn
status quo or by attempting to find and create new per se-is sufficient to higger intcrgm-ttlFcdmina-
differentiationsin its own favor, or both (seeChapter iioffi.r"
12). A more detailed specification of some of the a*GnG oim-eTl f an ouLgroupis sufficient
sEategiesand "solutiors " that can be adoptedin this to provoke intergroup compaitive or discriminatory
situation can be found in Tajfel (in press); we shall responseson Oe part of the in-group.
retnrn later to a discussionof some of them. For the In the initiat experimental paradigm (fajfel,
Fesent, it will be sufficient to statethat, whethervalid 1970;Tajfel et d., lg?l), thesubj€cb(borhchildren
or not, the hyp,othesisraisessomeimportant theoreti- and adul6 have acted as subjecb in the various
cal problems that necd to be considered. The first studi€s) are randonily classified as membersof two
questionis: what social-psychological procBSEs-arc nonoverlapping groups---astensiblyon the basis of
iffin- some tsivial performancecriterion. They then make
r@ "decisions," awarding
amountsof moneyto pairs of
u-ftFwnich-fi rrurrrdirerrft d6E6;;;ocial otlrer subjects(excluding self) in specially designed
booklea. The recipien8 are anonymous,except for
MC their individual code numben and their group mem-
quEsn'olls,
wemustnowabandon andcon-
speculation benhip (for example, membernumber 5l .of the X
sidersomerelevantdata. group and membernumber 33 of rhe I group). The
subjec6, who know their own group membership,
SOCIAL CATEGORIAT ION AND award thc amounts individually and aaonymously.
I NTERGRO UP DI SCRIM I NATION The responseformat of the bookletsdoesnot force th€
subjectsto act in termsofgroup membcrship.
The initial stimulus for the theorizing presented In this situation, there is neither a conllict of I
hcre was providcd by c€rtainexpcrimenal investiga- interess nor previously cxisting hostility betweenOre I
tions of intcrgroupbehavior.Thc laboratoryanalogof "groups." No social interactiontskesplace
bctwecn fl
rlzl-worl<i erinocenrsm is inGi6SiilGiGihe the subjecs, nor is th€re 8Dy rational link betwern ll
ctendencyto favor tle in-groupover the out-groupin economicself-inter€stand the strategyof in-groupt
An lntegrc ve Theory of lntergroup C;onflict 99
I
Thus, thcsegroupsare purelycognitive,
I favoritism. posaession r€porrs (Bilig, 1972; Turner,
and can be refened to asmininnl. llbjecs'
I 1975) shows that they do not shary any common con-
The basicandhighly reliablefinding is rhatrhe ception of the ',appropriate" or ..obvious,, way
fivial, ad hoc intergroup categorizationleads to in_ to
behave, that only a tiny minority have,o-"
goup favoritism and discrirnination against the out_ ido
of the hypothesis, and that this minority does not
group. Faimessis also an influential strategy.Thereis always conform to it.
alsoa good dealof evidencethat, within lhe pattem of The rnore general theoretical problem has been
r€spondingin termsof in-gmup favoritism, maximum refened to elsewhere by one ofus as follows:
difference(M.D.) is more importantto the subjects
than maximumin-group profit (M.I.p.). Thus, they Sjmply and briefly stared,rhe argument(e,g., Gerard
&
s€emto be competing with the out-gloup, rather than Hoyt, 1974)amountsto the following: thc subjectsacted
in
following a strategy of simple economic gain for terms of the inter$oup categorizstionprovidcd or imposcd
membersof the in-group. Other data from several by the cxperimentcn not neccssarilybectuse this bas bccn
exp€rimenb also show that the subjects' decisions suca€ssful in iDducingany genuincawarencssof member-
weresignificandy n€arerto the maximumjoint payoff ship in sepa.ate8nd distinct groups, but probably bccausc
(M.J.P.) point when these decisions applied to the they fclt that this kind of bchavic wasexpectcdof rhem by
division of money betwe€ntwo anonymousmcmbes thc expcriment rs, and ther€fore thcy conformcd to this
cxpectation. The filst question to ask is why should the
of the in-group than when they applied to two mem-
subjectsbc expe€tingthe cxp€rimentarsto crDe_.tof them
b€rsof theout-group;thatis, retativelylesswasgiven
thiskind ofbchavior?Thc Goard andHoy aruwerto this is
to the out-group, even when giving more would not that thc cxperimcntalrituatjon wasrigged aocausethis kind
havealfected the amountsfor the in-group. Billig & ol cxp€ctafonin thesubjects.This answerrrtainsib ptausi-
Tajfel (1973) have found the sameresults even when bility only if wc assumethatwhat wasno moreOtaoa hiot
the-assignment to groupswas madeexplicitly random. from thc cxp.riment€rs aboutthe notion of ,,groups,, being
This eliminated the similarity on the performance rclevant to the subjects' behavior had becn suffrcicnt ti
criterion witrin the in-group ss an alt€rnativeexplana- dctenninc, powcrful ly andconsistently, apanicularform of
tion of thc results(see Byrne, l97t). en expiiciOy lnt€rgroup behavi6. Il turn, if we assumc this___-and the
randomclassific{tiotr into groupsproved in this sOdy assumption is by no mcans uoreasonable_-wemust also
to bea morc potentdeterminantof discrimination than rssume dut this particulsr forr! of intetgroup behavior is
perc€ivedinterpersonalsimilarities and dissimilarities onc which is capableofbcitrg iuduccd by the cxpcrimetrters
much morc €asily than othcr forms (such as cooperation
not associatedwith categorizationinto groups.
betwcenthc groupsitr exto(ing tbe maximumtotaLamount
. Thequestion that arisesis whetherin_groupbias of money from the erpedmcntcts, or a fair division of the
in these "minimal,' situations is produced-by some spoils.between the groups,or simply |ardom rcspordi8g).
form of the cxperimenter effect or of the demand -by
And- this last assumptionmust bc backcd up in iL turn
charact€risticsof tle exp€rimentalsitration_in other aaothcr presupposition: aamcly, that for somc reaons
words, whedrerexplicit rcferenc€sto goup member- (whatever thcy may bc) competitive behaviq between
ship communicate to the subjects thit they are ex- groups, 8t least in out culture, is cxFaordinarily aasy
to
p€ctedto, or ought to, discriminate. The first Doint to Eigger off-at which point wc are back where we startcd
bc madcabout lhir interpreation of the resutt is that &om. Thc problcn thcn mustbcres&tedin ten|l! offhe occd
crplicit referencesto gmup menrbershipare logicalty to.spccify why 8 certain tr'nd of inlcrgroup bchavia can bc
nc.essaryfor opcrationalizing in thescminimal situa_ cllcrtcd so much Dore easily tha! otbcr linds: and this
specifrc€tioDis ccdainly rct Dadc if wc tlst contcnt with
tions the major independentvariablo--that is, social
thc cxplanatiotr that thc bchavior occurrpd bccauscit w6s
cat€gtrizationper se. This requircsnot merely that the
vcry casy for thc cxp€rimcntcG to makc it occur
subJectsp€rceivethemsclvesassimilar !o or different ffajfal,
inprcssl.
ftom others as individuals, but that they are mem-
bers of discrete and discontinuous cateeories--tlEt
Two points stand out: fint, minimal itrte.rgroup
is, "groups." Second, a detaited anal-ysisof the _.
discrimination is not based on incompatible group
Chaptet3

interests;second,the baselineconditionsfor inter- Socialgroups,undentoodin this sense,providetlrcir


group conpetition seemindeedso minirnal asto cause memberswith anidentificationofthemselvesin social
the suspicionthat we aredealingherewith somefactor terra. Theseidentificationsare to a very large extent
or proc€ssinherent in the intergroup situation ihelf. rclational and comparative:rhey define the individ-
Our theoretical orientation wasd€velopedinitially in ual as similar to or different from, as ..bett€r" or
rcsponseto theseclues from our earlier experiments. "worse"
than, membersof oher groups. (For a more
We shall not trace the history of its development, detaileddiscussion,see Tajfel, 1972b.) It is in a
however,but shalldescribeits prcsentform. stictly limited sense,arisingfrom theseconsidera-
tions,thatwe usethe te'.]l;.socialidentity.It consists,
SO CIAL IDENTITYAN D SO CIAL COM PARISON for the purposesof the presentdiscussion,of those
aspectsofaD individual'sself-imagethatderivefmm
Many orthodoxdefinitionsof "sociat groups,' the social categoriesto which he perc€iveshimself as
are unduly restrictive when applied to the context of belonging.With thislimitedconceptof socialidentity
intergmup relations. For example, when membersof in rnind, our argumentis basedon the following gen_
two national or ethnic categoriesinteract on the basis eralassumptions:
of their reciprocal beliefs about their respective 1. Individuals strive to maintain or enhanc€
categoriesand of the generalrelationsbetweenthem, their self-csteem: they strive for a positive self-
this is clearly intergroup behavior in the everyday concePt.
senseof the term. The "groups " to which the inter- 2 . Socialgroupsor categoriesandthe member_
actantsbelong need nat depqd upon the frequency ship of lhem are associatedwilh positive or negative
of intermember interaction, systems of role rela- value connotations. Hence, social identitv mav be
tionships, or interdependentgoals. From the social- positive or negativeacmrding to dre evaluaions
psychological pcrspective, the essential criteria for (which tendto be socially consensual,either within or
group membership, as they apply to large-scaleso- acrossgroups) of thosegroups that contribute to an
cial categories,arc that the individuals concemedde- individual'ssocialidentity.
fineliemselves and aredefined by othersasmembers 3. The evaluationof one's own group is deter_
ot a grcup. mined wi6 referenceto spe€ificother groups through
We can conc€ptualizea group, in this sense,asa social comparisonsin termsof value-ladenattibutes
collection of individuals who perceivc thernselvesto andcharacteristics. Positivelydiscrepantcomparisons
be memb€rsof the samesocial category,sharesome betweenin-group and out-g.oup produce high pres-
cmotional involvement in this common definition of tige; negatively disoepant comparisonsbetweenin-
themselves,and achieve some degrecof social con- group andouGgroupresultin low prestige.
sensusabouttlte evaluationof their group and of their From thescassumptions,somerelated theoreti-
membenhip of it. Following ftom this, our definition cal principlescanbe derived:
of intergroup behavior is basically identical to that of 1. Individuals srive to achieveor to maintain
Sherif(1966,p. 62): anybehaviordisplaycdby oneor positive socialidentity-
more actorstowsrd oneor moreothersthat is basedon 2. Positive social identity is based to a large
the actors' identification of themselvesand the others cxtent on favorable comparisonsthat can be made
asbelonging to different social categorics. bctweenthe in-group and somerelevatrtout-groups:
Social categorizations arc conccived here as the in-goup mustbe perccivedaspositivcly differcn-
cognitive tools that segment,clsssify, and cder the tiat€d or distinct from the relevantout-groups.
social environment, and thus enable.theindividual to 3. When social identity is unsatisfactory,indi-
undertakemany forms of social action. But they do viduals will strive cither to lcavetheir existing group
not mercly systematize the social world; they also ard join somemorepositivetydistinct group snd/orto
provide a systemoforientation for sefreference: they makc their existinggroup morepositively distinct.
crcate and define the individual's placc in society. The basic hypothesis, then, is that pressures
An lntegrativeTheoryof lntergroup Conflict
41
h€it to evaluate one's own group positively through throughsocialcomparison,whercasthe latter is based
rcial in-group/out-groupcomparisonslead social groups to on "realistic" self-intdest and representsembryonic
:tcnt attempt to differentiate themselvesfrom each other conflict. lncompatiblegroup goali are necessary
vid- (seeTajfel, 1974a,1974b;Tirrner,1975).Therearcat for
'or rcalisticcompetition,but mutualintergroupcompari_
least thlee classesof variablesthat should influence sons are neccssary,and often sufficient, for social
0ore intergroup differentiation in concrete social situa- competition.The latter point is consistentwith the
lna tions. First. individualsmust haveinternalizedtheir data from "minimal" group experimenbtlnt mere
lcra- group membershipas an aspe.t of thcil self.ccncept: awarenessof an out-group is sufficient to stimu-
ists, they must be subjectively id€ntified vith the rclevant late in-group favoritism, and the observations(Fer_
tose in-group. It is not enoughthat the othersdefine them guson& Kelley, 1964;Rabbie& Wi.lkens,l97l;
'rom
asa group, although consensualdefinitions by othe$ Doise & Weinberger, 1973) ttrat the possiUi ty oi
lf as can become, in the long run, one of the powerful social comparisongenera0es.,spontaneous,,inter_
otity causalfactors for a group's self-definition. Second, group competition.
gcn- the social situation must b€suchas to allow f6-iTfrter- Socialandrealistic competition alsodiffer in the
group comparisons that enable the selection and predictionsthat can be made about lhe @nseouenc€s
&nc€ evaluationof the relevant relational attributes. Not all for subsequent intergroupbehaviorof winninjor los_
sclf- betwe€n-groupdifferences have evaluative sisnifi_ ing.
-ftom realistic competition, the losing groups
cancc(tajfel, 1959), and those that do vary -After_
shouldbe hostileto the out-groupvicton, UotfrUe,-\
6er- group to group. Skin color, for instance,is apparently cause they have been deprived of a reward andJ
ive a more salient sttribute in the United State.sthan in becausetheir interactionhasbeenexclusively conllic_
tbe Hong Kong (Morland, 1969); whereas language tual. However, when winning and tosing establish
ions seernsto bean especiallysalientdimension ofseDarate shared group evaluations concerning comparative
in or identity in Rench Canada, Wates, and B jgium superiority and inferiority, then, so long as rhe terlr
Dan (seeGiles & Powesland, 1976;Fishman A Giles, in
of the competitionare perceivedaslegitimate and the
pr€ss).Third, in-groupsdo Dotcomparethemselves competitioniBelf as fair accordingto theselegitimate
?.tgr- with cvery cognitively available out_group:the out- teflns, the losing goup may acquiescein the superior-
)ugh group must be perceived as a rclevant comparison ity of the winning out-group. This acquiescenccby a
utes group. Similarity, proximity, and situational sa- group consideringi$elf aslegitimately . .inferior ,, has
sons lience are among the variables that detcrmine out- been shown in a rccent study by Caddick (1924).
rr1es- group comparability, and pressurestoward in_group
Severalother strdies reoort findings that arc in line
lm- distinctivenessshould increaseas a function of thii wilh this interpretation: losing in-groups do not al-
comparability. It is important to state at rhis point
ways d€rogate,but sometimesupgrade,their cvalua_
rcti- ttlat, in many social situations, comparability reaches tions of the wirming out-groups(tbr example, Wilson
a much wider range than a simply cooceived..simi_
& Miller, 196l; Bass& Dunteman,1963).
tain ladty " bctwe€nthe goups.
Retrosp€ctively, at least, the social_idetrtitv/
The aim of differentiation is to maintail or social-compar.ison theory is consistent with manv of
arge acldevesup€riority over an out-gmupon somedimen_
the studiesmentionedin tbe preccding seaion ofihi{-)
rade sroos.Ary suchact, th€refore, is essentiallycompai_
chapter.Itr partiorlar, in thc paradigm of the ..miiri_V
ups: tive. This conpetition requinesa situation of mutual
nul group" experinenb (such asTajfel et aJ.,lyll\,
$cn- comparisonand diffcrentiation on a sharcdvalue di-
&e intergoup discrimination can be conccivcd as
m€trsion. In theseconditions, intcrgroup ompctition
, being due not to conflict over monetarygains, but to ,
indi- wtuch may bc uDrelatcdto thc . .objectivc' ' goal rela_
differcntiations basedon camparisonsmadein terms:
roup tions bctwc€a the groups, catr bc pledic{cd to occur. of moneta4rrcwards. Money firnctioned as a dimcn_
0r to Timer (1975) has disringuishedberwe€nsocial and
sion of comparison(the only one available within the
.. instsumental,or ..realistic, " competition. The former
experimentaldesign), and the data suggestthat larger .
iures r3 motivated by self-evaluation and takes place
absolute gains that did not €stablish a differelcc in
42 ChapEr3

favor of the in-group werc sacrificedfor smallercom- The resultsshowedthat the Afis goups (H.I.)
parativegains, whenthe two kinds of gainsweremade weremore biasedthan the Sciencegroups(L:I.); that
to conflict. similar gmups differentiated more than dissimilar
There is further evidence Cfurner, in press-a) gI
that the social-competitive pattern of intergroup 49l9_9!gs9qano_!9m9gmgleventessso) m the Un_
behavior holds even when it conflicts with obvious stab& cogrtjgg and ttrat, onl6frE-6fthdFddlFs,
self-interest. In this study, the distribution of either therewasa signihaantmaineffectfor out-groupsimi- t
monelar5/rcwards or "points " was made, within the larity: in-groupbias increasedagainsta similar out- /
''minirnal" intergroupparadigm, group-Although thesedata are relativelycomplex,/
betwe€dselfandan
anonymous"other," who waseitherin the in-group they do supportsomeof our theoreticalexpectationsfl
or in &e out-group.As long as minimal conditions andprovidean illustation thatvariationsin in-groupr
existed for in-group identification, the subjectswere bias can be systematically predicted from lhe
preparedto give relatively lessto themselveswhenthe social-identity/social-comparison thmry.
award(eitherin points or in money) wasto be divided We havearguedthat social andrealistic compe-
betwecn self and an anonJ.mousmember of the il- tition are conceptualy distinct, although most often
group, ascomparedwith dividing with an anonymous they are empirically associatedin "real life." In an
memberofthe out-group. Theseresultsseemparticu- expaiment by Tumer and Brown (1976), an attempt
larly important, since the categoryof "self, " which was madeto isolatethe effectson intergroup behavior
is by no means "minimaf" or ad hoc, was set herc of the postrlatedautonomouspro@ssesattributedto a
against a tuly minimal in-gmup category, identical searchfor positive socialidentity. Ctrildren were used
to those used in the earlier experiments. Despite as subjects,and the manipulationsinvolved decisions
this stark asymmetry, the minimal gmup affiliadon by the subjectsaboutthe distributiotr of paymentsfor
affectedthe responses. participation in the experiment,to be sharedequally
The theoretical predictions were taken outside by the in-group, betweenthe in-group and out-grcups
of the minimal categorizationpdadigm in a further that were made rclevant or irelevant to compari-
study by Turner (in press-b). He used face-to-face sonswith the in-group's performance.Monetary self-
groupsworking on a disclssion task. In eac.hsession, interest (of a magtritudeprcviously asc€rtainedto be
two thre€-persongroups discussedan identical issue, of genuine significance to the subjec6) would have
supposcdlyto gain an assessment of thefuverbal intel- produced no difference in the distribution decisions
ligence; and then briefly comparedtheir respective involving the two kinds of out-group; it would also
performance. The subjects were lM male under- have led to decisionstending toward maximum in-
graduatcs.The criterion for intergroup differentiation group profit (M.I.P.) ratherthan toward maximum
was the magnitude of in-group bias shown in the difference(M.D,).
ratingsof thegoups' work. llalf &retriads, composed M.D. was the most influential strategy in the
of Arts students, believed that verbal intelligence was choices.Furthermore, whenthesubjecbcouldchoose
importart for them (High Importance, or H.I.); half, in-groupfavoritism(M.D. +M.I.P.) and/ora fairness
composedof Sciencestudents,did not (Low Impor- strategy,they wereboth morc discriminatory andless
tance,or L.I.). Half thesessions involvedtwo Arts or fair towardtherclevantlhan the irrelevantcomparison
two Sciencegroups (Similar Outgroup), and half in- group. Other measuresof in-group favoritism pro-
volved one Arts and one Sciencc group (Dissimilar ducedan interactionbetweenrcward level and typc of
Outgoup). Finatly, in rhe Stable Differencr cord! out-group: more discrimination against the relevant
tion, subjectswerc instructedthat Arts stsdens were than thc irrelevant goup with bigh revarG, and l€ss
defnitely sup€riorand Sciencestudentsdefinitely in- with low rewards. Whatev€f,may be other explaft-
ferior in verbal intelligencc; in the Unstable Differ- tions for this interaction,we can at lcast concludethat
ence condition, therc was no explicit statementihat when reward levels are morc meaningful, in-group
one c€tegory was better than the other. These vari- favoritism is cnhanccd against a more comparable
ablesweremanipulateAin a2xzxz factorial design. out-group, independendyof the group members'eco-
An lntegrativeTheoryol lnteercup @nflict
49
nomic interests.Indeed, insofar as the subiectsused indirectlyby rhe whole lireratureon racial
the M.D. strategy,they sacrific€d..objectivc" per- identifica_
tion and prefer€nce. The most important
sonaland group gain for the sakeof positive in_grbup f""tu." of
individualmobility is tiat thelow stahsof
distinctiveacss. one,sown
group is not thereby changed:it is an
On the whole, these studi€s provide some individualist
approachdesigned,at least in the shortrun, to achieve
confirmation for the basic social_identity/social_ a personal,not a group, solution. Thus, individual
comparison hypotheses. We shall now attemDt to mobility irnpliesa disidentificationwith the erstwhile
apply this analysisto somgofthe problemsraisedin in-group.
theeadier sectionsof this chapter.
Z
.l:r\lCreativity. The group membersmay
, positive
STATUS H I ERARCHI ES AN D SOCIALC HAN GE seet distinctivenessfor the in-groupby rede_
f.ning or altering the elements of thi comparative
The r€conc€ptualization of social status at- situation. This need not involve any changi
temptede{rlier in this chapterneedsnow to be made in the
group's ectual social position o( accessto
moreexplicit. Stafiisis not considercdhereas a scarce objective
resourcesin relation to the out-group. It is a group
resour€eor commodity, suchaspower or wealth: it is rather than an individualistic .fut"gy thut
the outcomc of intergroup comparison.It reflec8 a _.y
focus upon:
group's relative position on some evaluative dimen_ (a) C-omparingthein-group to thc outgroup
sions of comparison. Low subjective statusdo€s not on
somc-new dimension.Lemaine(1966)found, for
prcmoteint€rgmup competitiondircctly; its effects on ex_
ample, that childrcn'sgroupswhich could not com_
int€rgroup behavior are mediated by iocial identity pare themselvesfavorably with othen in
processes.The lower is a group's subjective status terms of
constructing a hut--because they had be€n
positionin rclationlo relevatrtcomparison assisned
groups,the poorcr building rnaterialslhatr the ouGg-up_t
lcss is the conribution it can makl to posltiveiocial id"d
seekout oth€r dimensionsof compaison invotvine
identity. Thevadery ofreactions to negadveor threat- . ,-to
new consfructionsin the hut's surroundings.
ened social idendty to be discussedbelow are an ThJ
problems that obviously arise here are dose of
elaboration of the principles outlined earlier in this legitimizing the valuc assigned to the new
chapt€r. social
produc6-f[st in the in-group and then in
the other
groups involved. To the extentthat this legitimization
l. In4ividual MobrTrry.The morean individual maythreatentheout-group'ssuperiordistinctiveness.
approachesthe structureof beliefs(mostofrcn socially an increasein intergroup tensioncanbe predict€d.
slnred) desoibed in the Introduction to this chapter
as (b! Changing the valuesassignedro the
that of ''social mobility, " the more it is likety that . attri-
-his he butes of the group, so that comparisonswhich.
will ry to leave, or dissociatehimself froln, were
".rr eS,evUystl negative are now perceived as positivc.
yll:.etllp. This usualty implies anempts,on an rne chsslc exampleis .,Black is beautiful.,,
ndividual basis,to achieveupwardsocialmobility, The
to salient dirnension---skin color_remains tfie
fuq a.Iower- to r higher-statusgroup. ln a same,
P* bur the prevailing value syst€m conceming
rour--goup hicrarchy, G. F. Ross (1975) fodd it is
a di_ andr€verscd.The sameproccssmay undertie
rcd. linear relationship betwcetr tow statJs :Je:y
and the Peabody's(1968) fnding that even when various
oes[ie to passupwardsinto anothergroup. Many
ear_ groupsagr€cabouttheir rospcctivecharacteristics
licr studies rcport the cxistencc of itong for&
for , the
upwad social movement in status hierarchies. trait is evaluatedmore positively by the group
rhat
Ten- possessesit.
denciesto dissociateonesclf psychologically
from (c) Changing the ou-Broup (or selectin*
reuow membersof low-prestigecategories the
are lqown out-g_roup)with which the in_groupis comparei_in
maDyoj us from everyday erperience; they
lo- have particul4r, ceasingor avoidhg to usethe iigh_status
oeon notedmoresystematically by Jahoda(196l) and our-group as a comparativc
A.lheb€f,8and Zavalloni (1969), among Jrame of relerence.
others,and rynerecornparisons are not madewith thehigh_status
44 ChapEr3

out-grcup, the relevant inferiority should decreasein unsatisfaclory social identity, to prcmote the wide-
salience,and self-esteemshouldrecover.Hyman's spreadadoptionof individual mobility strategies, or at
(1942)classicpaperon the psychologyof statussug- least initial attempGto makeuse of thesestrategies.
gestedthat discontentamong low-satus-group mem- Insofar as individual mobility implies disidenrifica-
bexsis lessenedto the degreehat intaclass ratherthan tion, it will tend to loosen the cohesivenessof the
intergroup comparisons are made. More recently, subordinategroup. This weakeningof subjective ar
Rosenbergand Simmons (1972) found that self- tadlment to the in-groupamongits memberswill tend:
esteem was higher among Blacts rylo made self- (a) to blur the perceptionof distinct group interests
comparisonswith other Blackggther than Whites. correspondingto the distinct group identity; and (b) to
Other work also suggests(seeI. Kae,, 1964;Lefcourt createobstaclesto mobilizing group membersfor col-
& Ladwig, 1965)that,in certaincircumstances, Black lectiveactionov€t their commoninterests.Thus. the
performancewas adversely affected by the low self- low morale that follows from negativesocial identity
esteeminducedby the presenceof the membersof the can set in motion disintegrativeprocessesthat, in the
dominantout-group. It follows that self-este€mcanbe long run, may hirder a changein the group status.
enhanced by comparing with other lower-sanrs Second,assumingthat the barrie$ (objective,
groupsrather than with thoseof higher sbnrs. This is mordfil'iEeological lrohibitions) to leavingone's
consistentwith the fact that competitionbetweensub- grcup arc strong, unsatisfactorysocial identity may
ordinate groups is sometimcs morc intensethan be- stimulate social creativity that tends to reduce the
twe€n subordinateand dominant groups-hence, for salietrc€of the subordinate/dominant groupconflict of
example,lower-classor "poor white" racism. int€rest. Stategy 2{c) mentionedaboveis likely to be
crucial hcf,esinc!, in general,accessto r€sourcessuch
3. Social Competition. The group members as housing,jobs, income,or educationis sufliciently
may seekpositive distinctivenessthroughdirect com- cetrtral to the fate of any group that the relevantcom-
petition with the out-group. They may try to revcrse parisons are not easily changed or devalued. Few
therelative positionsof the in- groupandthe out-group underprivileged groups would accept poverg/ as a
on salient dimensions. To the degreethat this may virtue, but it may appearmorc tolerable to the de-
involve comparisonsrelated to the social structure, it gee lhat comparisonsare made wilh even poorcr
implies changesin the groups' objective social loca- groups rather than with thosc drat are better off (see
tions. We can hypothesize, therefore, following the Runciman.1966).
R.C.T., that this stategy will generateconflict and As notedabove,somewriters(Festinger,1954;
antagonism between subordinate and dominant Kidder & Stewart,1975)imply that sFategy2(c) is a
groups insofar as it focuseson the distibution of dominant response to status differences between
scarceresourc€s.Data relevant to this strategy have groups. The assumptionis that intergroupcomparabil-
beenreferredto earlier in this chapter. ity decreases asa directfunctionof perceiveddissimi-
lrt us assumeas an ideal casesomestratifica- ladty. If this were the whole story, then, somewhat
tion of social groups in which the social hierarchy is paradoxically, lhe creaiionof a consensualstatussys-
reasonablyconelated with an unequaldivision of ob. tem would protectsocial identity froin invidious com-
jective resourcesand a correspondingstatus system parisons.The causalsequenceworild bc as follows:
(basedon tle outcomesof comparisonsin t€rms of similar groupscomparewith cachotheq he outcome
thoseresources).Under what conditions will this nrt dercrmincstheir relativeprestige;the perceivedstatus
lead to intergroup conflict----or,more precisely, to the difference r€duces their .similarity and hence com-
developmentof competitive ethnoceutrismon the part parability; intergroup comparisonsceasbto be made;
of the subordinategroup? subjective superiority and inferiority decreasein sa-
First, to the extent that the obieaive and the lience; mrrespondingly, the groups' respectiveself-
subjeEiild prohibitions to "passing"- are weak (see este€msretlm to theL original point. There may be
our earlierdisclssionof the "social rnobility" stuc-
c)o,tn"
ture of beliefs), low status may tend, in conditions of
occasionswhen this social-psychologicalrecipe for
the maintenanceof the statusquo can be obeervedin
An lnEgrative Theoryof tnteryroup Conflict

somethinglike its pureform. However,we shall areue tions 8re p€rc€ivedas immuable, a part
prescntlythat thereare manyshtus diffcrencesdraido of the fixed
order of things, then social identity is
not reducecomparability. secure..Itbe_
comes inseclre when the existing state
For the moment, we can note that bodr indi- of affairs be_
b be.erystio1ed.An important coroltary
vidual mobility and someforms of socialcreativity llm to this
argumentis that the dominant or high_status
can work to r€duce intergroup conflict over scarce groups,
too, can experienceinsecuresocia
resourc€s-though with different implications. The iOentit!. eny
former is destructiverf subordinate-groupsolidarity gT., . dt. superior
9i"!i":,jvely position
,rl groui,
lmpt.iesa polential loss of positivecomparisoniani
andprovides no antidoteto.negativesocial identiw at posslble negative comparisons,which
grolp level, The lattermayrestorcor creat!a posiive must be
guarded against. Sudr a threat may
self-imagebut, it can be surmised,at thepriceeidrer derive from the
a:qvjry of rhe low-statusgroup or from
of a collective repressionof objectivc deprivation or, a conflict
*".high_sratus. gmup,s own value sysrem
perhaps_, of spuriousrivalry with someotherdeprived :1Tn
(lor rhe sociopoliticalmorality) ani he
group.Ilis in tlds context that :nmple,
actual foundationsof its sup€riority. Like low_status
Canadians,
groups,the high-statusgroupswill react
more to insecue
social by searchingfor enhancedgroup
(see _idetrtity
q$uncuveness -
In brief, then, it is tue that clear_cut
By revming the conditions underwhich social sta s
. maf ro a quiesc€ntsocial sysremin
stratificatioo do€snot produceintergroupconflict, we !ifl11ence1. .leaa
whrch neitter the ..inferior', nor the .,superior,,
c:n hypothesiz€that negativesocial identity promotes
groupswill showmuchethnocentism.But
subordinategroup competitivenesstoward the dom_ this ..ideal
type" situation must be consideredin relation
tnatrt group to the degre€that (a) subjective to tre
iden- p€rc€ivedshbility and tegitinacy of the
ufi:1tr:l wi_rhrhe subordinaregroup is mainrained; system.per_
c€tvd luegittmacy and7orinstability provide
alld O) the dominart group continuesor beginsto new di_
be mensionsof comparability Oat aredirectly
a relevant comparisonFoup. As a grear relevantto
fTi:"d I me atutudes and belravior of the social groups
deal of work has becn done in social psycholog-y in-
on volved, whatevertheir positionin dresystem..ibis
the determinantsof cohesivcnessand loyalty iirhin is
th€social-psychologicalcouDterpart!o what is
groups,we shall conc€ntrateon the seconi widely
condition. known today as .,the revolutionof rising expecta-
let us consider a comparison betwe€ntwo
foofball tions. ' ' Providing that individual mobitity is unavail_
teamsthat have come first and secondin their
league, able or undesirable,consensualinferiority wilG
respectively. There is no argumentabout wh.ich
bas rcjected rnostfapidly when the situation is perceived
the bigher stafirs, but alternative corparativc
out- botr u:srabl-e. illegitimate. This is'prcbabiy
cojncswere 8nd, h the futue, still will be possible. T Td
lynen the new seasonbegins,the teams the s€t of conditions underlying the devilopncnt
wh be as ot ethnocentrism among Black Americans,
comparatle and competitive as they hadbeen French
bcfore. Lanadrans,and New Zaland Maoris, for
rn |hrs rnshnc€, tle saoJs differencc instance.
does not re_ V.aughan(in prcss) reporb that the perceivcd
ducc-the meaningfulnessof comparisons feas!
becausei, Durry ot social change (probably including,
can be changed. in this
usr'4c€, tbc pcrccivcd illegitimacy of thi prcscnt
This cxampleilusbares Tajfel,s (1974a)distinc-
-. .
uon Dctwe€nsecrra andirsrcure itrle{group i, an important pt€dictor of fte deveJoping'
compari- lltual*).
Maori cftnocentrism; N. Friedman (1969)
sons.fte crudilEid?G-6-G?Ftinction ariui
is whcther *F1 wc nry tenn the ..cognitive alteiaa_
c:g:*-:h?natives to rhe actual ourcomc arc 9il
avarraDtg-wh€thtr other outcomes tive" of Black nationalism in the de-vclopingcoun_
arc conccivable. dT
Jtatus diffe(ences betwccn social groups ,*". influentiat in cnhancingBlack Aierican
in social socialidentitv.
s)ttcms showitrg various degrecs
of stradficationcan On the other hand, wher the dominantgrcup
oe oEttnguishedin the sameway. Whcrc _ or
statusrrla_ seclionsof it perceive their superiority as teg'itir;b,
46 Chapter3

they will pmbably react in an intenselydiscriminatory and empiricallydistinctfrom out-grouphostility. On


fashion to any attempt by the subordinaF group to the otherhand, social-identityprocessesmay provide
changethe intergroupsituation. Suchpedrapswasthe a sourceofintef,groupconflict (in additionto the cases
post-bellum sinration in the Southern United States: outlined above)to the degee that the groups develop
the Whites, threatenedby those who had been their conflicting interestswith respectto themaintenanceof
slaves, rapidly abandoned their Patemalistic ste- the comparativesituation asa whole. It seemsplausi-
"childlike" in favor of ble to hypothesizethat, when a group's action for
reotypesof the Blacks as
openly hostile and derogaory ones (see Van der positivedistinctiveness is frustrated,impeded,or in
Berghe, 1967). The reactionsof illegitimately su- any way activelypreventedby an out-group,this will
perior'groups arc more complex (I\rmer & Brown, promote overt conflict and hostility between the
1976). It se€msthat conflicts of valuesarereducedby groups.This prediction,like many others, still re-
greaterdiscriminationwhensuperiority is assued, but mainsto be tested.
by lessdiscrimination whenit is unstable.This calls to "O BJECTIVE' AN D'SU BJ ECTIVE"
mind somePrisonerDilemma surdiesin which White
CONFLICIS
discrimination againstBlack opponentsincreasedthe
more cooperativewasthe opponent,but decreasedthe None of the argumentsoutlined in this chapt€r
more comptitive he was (Baxter, 1973; Cederblom must be understoodas implying.that the social-
& Dien, 19?0). Baxter suggestedin the tide of his psychological or "subjective" t)?e of conflict is
") being consideredherc as having priority or a more
article ("Prejudiced Liberals? that a conflict of
valuesmay underliehis data. important causal function in social rcality than the
"objective'' determinantsof social conflict of which
Many of the points and hypotheseswe have
advancedin this chapter are not, in thernselves' new the basic analysismust be soughtin the social, eco- ,
(see,for instance,M. Sherif,1966;Runciman,1966; nomic, political, and hisorical strucnrresof a soci-
Milner, l9?5; Billig, L976).What is trew,we think, ety. The major aim of the presentdiscussionhasbe€n
is the integration of the three processesof social to determine what arc the points of insertion of
categorization, self-evaluation through social iden- social-psychologicalvariablesinto the causal spiral;
tity, and inte{grcup social comparison,into a coher- and its argumenthas becn that, just as the effects of
ent and testable framework for contributing to the thcsevariablesare powerfully determinedby the pre-
explanationof various forms of intergroup behavior, vious social, economic,and political pmc€sses,so
social conllict, and social change. This framework they may also acquire, in turn, an autonomousfunc-
contains possibilities of further development, and, tion that enablesthem to deflect in one direction or
to this extent, we hope that it may stimulate theo- anotherthesubsequentfunctioningof theseprocesses.
retically dirccted researchin areasthat haYenot be€n It is rle-adyimpossible itr most natural social
consideredhere. situations to distinguish betweendiscriminatory in-
But some cautionarypoins should be made. tergroup behavior based on real or perceived con-
The equation of social competition and intergroup flict of "objective" interests betwe€n the gmups
conflict madeaboverestson thc assumptionsconc€rn- and disgimination based on attempts to establish
ing an "ideal O?e" of sociat statification in which a positively-valued distinctivene,ssfor one's own
the salientdimensions of intergrouPdiff€rentiation are group. However, as wc have argued,the two can be
those involving scarceresources.In this resPect,we distinguishedtheoretically, since the goals of actions
have simply bonowed the central tenet of the R.C.T. aimed at the achievementof positively-valued in-
Thcre is no reason,in fact, to assumethat intergroup group distinctivenessoften relain no value outside of
diffe.rentiationis inhcrently conflictual. Some of the the context of intergrouPcomparisons.An examPle
experimentalwork that is proceedingat pr€scntat the would be a group that does not necessarilywish to
University of Bristol points to the conclusion that increasethelevel of its own salariesbut actsto prevent
eyaluativederogationof an out-group is conceptually other groupsfrom getting nearerto Ois level so that
An lntegntive Theoryof tntergroup Conflict 47

differentialsarenot eroded. B ut the difficulty w ith this types and all other aspects of the ..irrelevant"
example---aswith many other similar examples-is in-group/out-groupdifferentiationsso well described,
that, in this case,the preservationof salary differen- for example, in Sherift s[rdies. The fmt categoryof
tials is probably associatedwith all kinds of "objec- actions is both commonsensicallyand theoretically
tive" advantagesthat cannot be defined in terms of ac@untedfor by assumingnothing more than the
moneyalone. In turn, .rone of theseadvantageswill group's desire to win the competition-altfrough this
againmakesCSg only in lte comparativeframework posesall the theoretical"comparison" problemsdis-
of intergroup -ompetition. Despite this confusing cussedin this chapter;the secondcategoryof actions
network of mutual feedbacks and interactions. the can be directly and parsimoniously accountedfor in
distinctions made here are important becausethey terms of the social-comparison-social-identity_
help us to understandsome aspccb of int€rgroup be- positiveingoupdistinctivenesssequencedescribed
havior which haveoften beenneglectedin the past. here.
A fifther distinction must be made betwe€n The "implicit" conflicb are tlose that can be
"explicit" and "implicit" conflicts-a
distinction shown to exist despite the absenceof explicit in-
that has to do with objectivity in a different way. A stitutionalization or even an informal normative ac-
conflict can be "objective" despite the fact that the c€ptanc€of their existenceby the groups involved.
goalsthe groupsare aiming for haveno valueoutside The proof of their existenceis to be found in the large
of the contq(t of intergoup comparison. This is so number of studies (and also everydayoccurrencesin
when the conflict is instiqrdonalized and legitimized " teal life')
in which diff€rentiations of all kinds are
by rules and Dofms, whatev€r their origin, that are madebetweengroups by their membersalthough, on
acccptedby the groups involved. This was the casein the face of it, there are no "r€asons" for thesedif-
Sherif's studi€sin their phaseof competition betwecn ferentiations to occur. Clear examplesof this have
lhe groups;andit alsois the casein any footbau match been provided in several studies mentionod in this
atrd in countlessoth€r social activities. The behavior chapt€r in which the intoduction by the subjecn of
toward out-groupsin this kind of cotrflict can be clas- various intergroup differentiationsdirectly decreased
sificd, in tum, into two categori€s,one of which can the "objective" rewards that could otherwisc have
bc rcfcrre;d to I instrununtal and the other as been gained by the in-group, or evendirectly by the
noninstrumental.The instrumental category consists individual. Findings of this kind, which can be
of all thoseactioDswhoseexplicit aim can be directly generalizedwidely to many natural social situations,
rclated to crusing the group to win the competition. provide a clear example of the n€edto introduc€into
Thc noninstrumcntatcategory, which could be re- the complex spiral of social causation the social-
ferred to as "gratuitous" discrimination against the psychological variabtes of the ..rclational " and
out-goup, includes the creation of negativc stereo- "cornparative "
kind discussedin this chapter.

You might also like