manuscripts 3 social identity, social categorization and social comparison in intergroup behaviour

Upload: raul-lleal

Post on 03-Jun-2018

235 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    1/64

    SOCIAL IDENTIT Y, SOC IAL CATE GORI ZA TION AND SOCIAL

    COMPARISON IN INTERGROUP BEHAVIOUR.

    I . The t heoret ical and r esear ch backgr ound

    The pr esent r ese ar ch proposal is not base d on a theor y whi ch is fi rml y

    and fi nall y art i cul ated; it is not, howev er , an ent ir ely new r esear ch depart ur e.

    For some three or four year s t here has bee n an inter l ocking deve l opment , in

    Bri stol and el s ewher e, of experi ment al s t udies and t heoret ical i deas rel ated

    t o t hem, many of whic h wer e defi nit el y post hoc. The aim of this proposal is

    to enable us to continue this process with the hope that, at the end of the

    proj ect , we shall have a more sol id basi s for a r espect abl e t heory.

    Soci al psychol ogy of intergroup rel ati ons has not fared t oo wel l in the

    l ast thirt y year s or so. Char act er i st i cal l y, in their chapt er on "Gr oup pr oce s s e

    in t he Annual Rev i ew of Ps y chol og y , Ger ar d and Mi l l er (1967) devot ed t o it a

    dozen l i nes or so in a text of about 40 pag es . The i mpact of t hese f ew l i nes

    i s: (a) that i ndeed ver y l it tl e experi ment al wor k on intergr oup pr oce s se s is

    being done; and (b) t hat t hi s i s due t o t he met hodol ogi cal dif fi cul t i es of

    creating intergroup situations in the laboratory.

    The major t r ends can be bri efl y charact eri zed as fol l ows devel opment s

    from vari ous ver si ons of t he fr ust rat i on- aggression hypot hesi s wel l r epr esented,

    for exa mpl e, by t he work of Berkowi t z ( e .g . 1962, 1965); and t he quasi-

    experi ment al st udi es of gr oups in confl ict deri ving from t he work of Sheri f

    (e .g . 1 9 6 6 ) . In addi t i on, t here have been s ome i sol at ed st udi es about t he

    eval uat i on of i ngr oups and out groups ( e .g . Bas s & Dunt emann, 1963; Doi s e ,

    1969; Ferg uson & Kel l ey , 1964; Pea body , 1968; Rabbi e & Hor wi t z, 1969;

    Rabbie & Wi l k e ns , 1971; Wi l s on & Kat ayani, 1968; Wil son & Mi l l er , 1960)

    whi c h t og et her , ho wev er , di d not amount t o a cl ear - cut t heor et i cal adv anc e

    in t he fi el d. The ri ch t radit ion of t he wor k on prej udi ce init i ated soon after

    t he war i n t he Uni t ed St ates (Adorno et al . 1950) was mai nl y, if not exc l usi ve l y,

    focused on the genesis of hostile and discriminatory attitudes and behaviour;

    t he s ame is t rue of t he wor k on st ereot ypes and ot her cogni t i ve aspect s of

    pr ej udi ce, al t hough t here have bee n sc mo at t empt s to r el ate this wor k to more

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    2/64

    a

    general aspect s of cogni t i ve funct ioning (e .g . Al l port , 1954; Campbel l , 1967;

    Taj f el , 1959; Tajf el , 1969a).

    Independent l y of the detai l of the vari ous approaches t o the psy chol ogy

    of int ergroup r el at i ons, one gener al st atement about t hem can be made whi ch

    does not (as it appear s to t he present wri ter) do t hem i nj ust i ce. They f ocused

    on intergroup processes taking as their point of departure attitudes and

    behavi our t owar ds t he out gr oups. In some c as es , as in t he wor k on fr ust rat ion

    and aggression, these outgroup attitudes and behaviour were seen as an

    i ns t ance of a more gener al l aw of human mot i vat i on, i n some s ens e pr el i minary

    to any social cont ext (cf . Taj f el , 1972a). In ot her s, such as the work of

    Sheri f , t he emphas i s has been on t he devel opment of out group at t i t udes, and

    al most epi phenomenal l y , of ingroup aff i l i at i ons. Both t he ingroup and out group

    att it udes wer e s een as a result of emergi ng soci al nor ms, dir ect l y due to an

    expl ici t intergroup confl ict of goal s. This conflict was not only clearly explicit

    to t he subj ect s; it was al s o clearl y defi ned as such by t he experi ment ers ; t he

    creation of it and the study of its results were the aims of the research.

    Muc h of the argument t o fol l ow can be sai d t o be bas ed on a si mpl e

    st atement : i n order for t he members of an ingr oup to be abl e to hat e or di sl i ke

    an outgroup, or to discriminate against it , they must first have acquired a

    se nse of bel ongi ng t o a group whic h is cl earl y dis t inct f rom t he one t hey hat e,

    di sl i ke , or discr imi nat e agai nst . Muc h of t he t radit ion in t he l it erature (not

    onl y i n social psychol ogy) ascr i bes the acqui si t i on of this sense of bel onging

    to t he exi st ence of out groups per cei ved as t hr eat s, common enemi es , et c. A

    weak er pr oposi t i on i n t he s ame t r adi t i on i s t hat t he ex i s t enc e of s uc h out gr oups

    at l east cont ri but es t o, or incr eas es , t he intensit y of ingroup aff i l iat ions. The

    exi st ence or st rength of the i ngr oup are t hus se en as phenomena deri ved from

    t he r el at i ons bet wee n t he ingroup and it s outgr oups. In some cases, this

    consists of presumed intragroup effects of various kinds of direct or projected

    out group hosti l it y; in ot hers, it is seen as a dir ect result of an "obje ct i ve"

    confl ict of inter ests bet ween t he gr oups. The emphasi s r emai ns t he s ame.

    Ther e i s no doubt t hat t hi s out group- i ngroup s equence of at t i t udes and

    behavi our has a great deal of val i dit y, both intuit ivel y and as a result of a

    great mass of empir ical ev i dence. But t he emphasi s- under st andabl e as it has

    - 2 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    3/64

    been, mainl y for social r easons - is too one- si ded. An adequat e social

    psy chol ogi cal t heor y of int ergroup behavi our must t ake i nt o account bot h

    causal di rect ions: from ingroup proc es ses t o outgr oup behavi our and att it udes

    as wel l as t he opposi t e one whic h has bee n unti l now t he pr incipal obj ect of

    t heory and r es ear ch. Even if it is t rue t hat origi nall y many groups are cr eat ed

    as a common shel t er for t heir member s f rom outs i de t hr eats a nd danger s

    (human or not ), it i s equal l y tr ue t hat in any c ompl ex s oci et y an i ndivi dual

    confr ont s from t t l a beginni ng of hi s li fe a complex net work of groupings

    whi c h pr es ent s hi m wit h a net wor k of r el at i ons hi ps i nt o whi c h he mus t fit

    hims el f . One of t he most import ant a nd dur abl e pr obl ems t hat i s pos ed to an

    i ndi vi dual by his i nsert ion int o societ y i s to f i nd, create and defi ne his pl acein t hese net wor ks . It is r easonabl e to as s ume t hat both his i ngroup and

    out group att i t udes and behavi our must be det er mi ned, t o some ext ent at l eas t ,

    by t his cont inuing process of self - definit ion.

    An ear l y v er s i on of t hes e i de as l ed t o t he f i rst exper i ment s we conduct ed

    in Bri st ol s ome t hr ee or four year s ag o (Tajf el , 1970a; Tajf el et al . 1971).

    Their aim was to est abl i sh minimal condit i ons in whic h an indi vi dual wi l l , in

    hi s behavi our , di st i nguis h bet ween an i ngr oup and an out gr oup. In order t o

    cr eate such mi ni mal condit i ons we at t empt ed t o eli mi nat e fr om t he experi ment al

    si t uati ons al l t he var i abl es t hat normall y l ead to ingroup favour it is m or

    di scr i mi nat i on agai nst t he out group: face- t o- face interact ion; confl ict of

    interest s; any possibil it y of pr evious hosti l it y ; any "ut il it ari an" or inst ru

    ment al l ink bet ween t he subj ect s' r esponses and their sel f- int erest . In

    addi t i on, we enabl ed t he subj ect s t o chose amongst a vari ety of st rat egies in

    their r es ponse s , some of whic h wer e more "t Rt i oaal '1 or "us e f ul 1' t han cr eat i ng

    a dif fer ent iat ion bet ween t he gr oups. The s ubject s first perf ormed a rel ati vel y

    tr ivial t ask (guess i ng number s of dots i n rapidl y pr oject ed cl ust er s, or

    expressing preference for the paintings of one of two fairly abstract painters,

    Klee and Kandinsky). They t hen wor ked separat el y in i ndi vi dual cubi cl es .

    Thei r t ask was t o deci de (on a number of payment matr i ces) about div i si on of

    point s wort h money bet ween t wo ot her s ubj ect s . They k new what was t heir

    own group members hip (under- or over- est i mati on of dot s; or pref er ence for

    one or t he other pai nt er ), and the gr oup member shi p of thos e bet wee n whom

    - 3 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    4/64

    I *

    %

    t hey wer e di vi ding t he money; but t hose other s wer e des i gnat ed by code

    numbers , and t heir ident it y was unknown. The result s wer e very highly

    si gni fi cant i n t he di r ect i on of awar ding more money t o members of the "i ngr oup"

    In t he s econd set of exper i ment s, t he mat r i ces wer e s o const r uct ed that we

    coul d as s es s t he separate "pul l " of several var iables on the deci si ons. These

    var i abl es wer e: maxi mum joint profit ( i .e . t he str ategy of awar ding t he

    maxi mum joi nt amount on eac h mat r i x, so that al l t he s ubj ect s t ogether - who

    knew each other wel l before t he experi ment s - coul d get t he greatest possi bl e

    amount of money out of t he exper i ment er ; max i mum pr ofi t for member s of the

    i ngr oup; maxi mum dif f er ence i n favour of t he ingr oup at t he pri ce of s acri fi ci ng

    both t he above advant ages ; and fai r ness of choi ces . Of these var i ables , t he

    first - maximum joint profit - exerted hardly any pull on the decisions;

    maxi mum i ngr oup profi t wa s i mpor t ant , but somet i mes not near l y a s i mport ant

    as achiev i ng maxi mum di f fer ence in favour of t he i ngr oup. Fair ness was al so

    a significant variable and served to moderate the excesses of ingroup

    favourit ism.

    Two si mpl e and overl appi ng expl anat i ons are avai l abl e to account for

    t hese r esult s, a "nor mat i ve" one and a "l ear ni ng" one. The fi rst i s t hat ourschool boy subj ect s , a ged 15 to 16 ye ar s , s aw t he si t uati on as one of "t eam

    compet i t i on" in whi ch one s houl d make one's own team wi n at what ever cos t .

    The second, that - in a new situation - they engaged in ingroup behaviour

    whi c h had be en r ei nf or ced on count l es s oc c as i ons i n t he pa s t . Bot h t hes e

    expl anat i ons are sensi bl e; t hey are al so quit e "uni nt er est i ng" - uni nt erest ing

    becaus e not genui nel y heur i st i c. If our subj ect s had chosen str ategies of

    choices leading to maximum joint profit , the same explanations could still

    s er v e, in one form or anot her. If t hey had chose n onl y the st rat egy of f air ness

    wi t hout t hat of i ngr oup f av our i t i s m, one coul d st i l l "ex pl ai n ' t hei r r es po ns es

    st arti ng fr om norms and previ ous r ei nf or cement s. My argument i s not that

    t hese expl anat i ons are inval i d. It i s rat her t hat , i n addit i on to their capaci t y

    to expl ain indis cri minat ely al l kinds of r esul t s, they are at a l evel of generali t y

    whi c h pr event s t hem f rom ser vi ng as a poi nt of depar t ur e f or ne w and mor e

    searching insights about intergroup processes.

    - 4 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    5/64

    It i s t he choice by t he s ubject s of t hese part icul ar norms bas ed on t hese

    particular reinforcements which defines the problem and provides a departure

    for some r esear ch quest i ons about t he psy chol ogy of i nt ergroup r el at i ons.

    This is particularly so in view of the fact that the results have since beenr epl i cated in sever al experi ment s both in Bri tain (Bi l l i g, 1972; Bil l ig & Taj f el ,

    1973; Tajf el & Bil l i g, 1973; Tur ner , 1973b) and el sewher e (Deut s ch et al . ,

    1971; Doi s e , et al . . 1972; Sol e et a l . , 1973).

    The probl ems of an i ndiv i dual 's self - definit ion in a s ocial cont ext ,

    bri efl y ment i oned abo v e, can be rest ated in t erms of the noti on of s ocial

    ident it y. Wo need to postulat e t hat , at l east i n our ki nds of s ociet ies -, an

    i ndiv idual st ri ves to achiev e a sat i sfactory concept or image of hi msel f . This

    wa s one of t he ba s e s of Fe s t i ng er 's t heor y of s oc i al compar i s on (1954).

    Fest inger, however, was almost exclusively concerned with social comparison

    made bet ween i ndi vi dual s and wi t h eval uat i ons of onesel f and others made by

    means of t hese inter- indivi dual compari sons. This int er- indivi dual emphasi s

    negl ect s an important cont ri but ing aspect of an indi vi dual 's sel f- definit ion:

    the fact that he is a member of numerous social groups and that this member

    ship contributes, posit ively or negatively, to the image that he has of himself

    Four l i nked concepts wi l l be empl oyed in order to proc eed wi t h t hi s

    dis cus si on. They are: social cat egor izat ion, social i dent it y, social compari

    s on and^ps y c h ol og ic al di s t in c t iv enes s . ) * ^ cflet* *

    The process of categorization, as it is used by the human individual in

    order to syst emati BO and simpl if y his envi r onment , pres ent s cert ain t heoret ic

    cont i nuit ies fr om t he rol e pl ayed by categor i zing i n per cept ual act i vi t i es t o

    it s r ole in t he orderi ng of one's soci al envi r onment . For our pur pose , soci al

    cat egor i zati on can be under st ood as t he orderi ng of s oci al envi r onment in t erm

    of s ocial cat egor i es, t hat i s of groupi ngs of pers ons i n a manner whic h is

    meani ngf ul t o t he subje ct . Ther ef or e, i n our dis cus s i on t he ter m gr oup"

    denot es a cogni t i ve enti ty that i s meani ngf ul t o t he subj ect at a part icul ar

    poi nt Of t i me and must be di st i ngui shed fr om t he way i n whi ch t he t erm "gr oup

    is used in much of t he social psyc hol ogi cal Ut erature where it denot es an

    "obje ct i ve" (most oft en face- t o- face) r el ati onship bet ween a number of peopl e

    In other wor ds , social categori zat ion is a proces s of bri nging t ogether s ocial

    - 5 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    6/64

    obj ects or events in gr oups whic h ar e equi val ent wit h regard to an i ndiv i dual 's

    act i ons, i nt ent ions, att it udes and syst ems of bel i ef s .

    The second concept we must introduce here is that of social identity.

    For our pur poses we shal l unders t and soci al i dent it y as t hat part of an

    i ndiv i dual 's sel f- concept whic h der i ves f rom hi s knowl edge of hi s member ship

    of a social group (or groups) together with the emotional significance attached

    to that membership.

    Social categorization must therefore be considered as a system of orien

    t ati on whic h cr eates and defi nes the i ndiv i dual 's own pl ace in soci et y. As

    Berger (1966) wr ote: "Ever y soci et y cont ai ns a r epert oir e of i dent i t i es that is

    part of the "obj ect i ve k nowl edge" of it s member s '1 (p .106). "Sociot y not onl y

    defi nes but cr eat es psychol ogical r eal it y . The indiv idual r eal i ses himsel f in

    society - t hat i s , he r ecognizes his ident it y in social l y def ined t erms and

    these definit ions become reality as he l ives in society" (p.107).

    Several consequences regarding group membership follow upon this

    "r ecogni t ion of ident it y in social l y defi ned t er ms ". They can be desc r i bed

    as f ol low s :

    a) It can be as s umed t hat an indi vi dual wi l l t end t o r emai n a member of

    a group and se ek members hip of new gr oups i f t hese gr oups have some cont ri

    but ion to make to t he posi t iv e aspect s of his social ident it y; i .e . to t hose

    aspec t s of it fr om whic h he der i ves some sat i sf act i on.

    b) If a group does not sat i sf y t hi s r equi r ement , t he i ndiv i dual wi l l t end

    to leave it unless

    (i) l eavi ng t he group is i mposs i ble for some "object i ve"r easons or,

    (ii) it confl i ct s wi t h i mport ant val ues whi ch are t hems el ves a part of

    h is ac c ep t able s oc ia l iden t i t y . ,

    c) If l eavi ng the group pr esent s t he di ff icult i es j ust ment i oned, t hen atleast two solutions are possible:

    (i) t o change one's i nt erpretat ions of t he att ri but es of t he group so

    that it s unwel come features (e .g . l ow status) are either j ust i fi ed or

    made accept abl e t hrough a reinterpret ati on;

    (ii) t o accept t he si t uati on for what it is and engage i n social acti on

    whi c h woul d l ead t o des i r abl e c ha ng es i n t he s i t uat i on. (Of c our s e,

    t here may be var i ous combinati ons of (i) and (ii) , s uch a s , for

    ex ampl e, when t he negat i ve att ri but es are just if ied and social act i on

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    7/64

    d) No gr oup l i ves al one - all gr oups i n soci et y l i ve in t he mi dst of ot

    gr oups. In other wor ds, t he "posi t i ve aspect s of soci al i dent i t y in (a) abov e,

    and t he rei nt erpretat ion of att ri but es and engagement in s oci al act i on in (c)

    abo v e, onl y acqui r e meaning in r el ati on t o, or in compar i sons wi t h, other

    groups.

    It is this comparative perspective that links social categorizing with

    social i dent it y. In his t heory of social compari son pr oce s se s, Fest inger

    (1954) hypot hesi zed t hat "t her e exi s t s , i n t he human or gani sm, a dr i ve to

    eval uat e his opini ons and his abil i t i es ". Hi s sec ond major hypot hesi s in t he

    same paper was that "t o t he extent object i ve, non- social means are not

    ava i l abl e, people evaluat e their opinions and abil it ies by compari son r espec

    t ivel y wi t h t he qpi ni ons and abi l i t ies of ot her s ". But there are s ome dif fi cul t ie

    wit h t he conc ept i on t hat s oc i al compar i s ons onl y t ak e pl ac e "t o t he ex t ent t hat

    obj ecti ve non- social means are not avai l abl e". Fest inger's exampl e is that

    "one , coul d, of cour s e, t est t he opi ai on t hat an obj ect, was f r agHe by hit t ing

    it wi t h a hammer ". I can confi rm t he opi ni on t hat a bed is for l yi ng- down- on

    by lying down on it until I discover that this particular bed in this particular

    room of the cast l e bel onged t o t he Duke of Ur bino and is most defi ni t el y notfor l yi ng- down- on. Very oft en, t he "obj ect i ve non- soci al means that may

    appear t o an obser ver to be avai l abl e for the test ing of opini ons do not have

    much val idit y unl ess t hey are used i n conj unct ion wit h t he s i gnif i cance that

    t hey acquir e in their social set t ing. The cas es whic h l ie out si de t his r ange

    are usual l y tr ivial in t he anal ys i s of social behavi our . In addit i on, social

    real it y can be as "obj ect i ve" as i s non- social real it y , and convers el y "obj ec

    t i v i t y " c an be as "s oc i a l 1 as it is "phy s i ca l ". In some cul t ur es, thunder and

    lightning are as indisputably signs of anger of supernatural powers as they are

    bursts of sound and light.

    The criterion of "objectivity" cannot be based on classifying phenomena

    as bei ng of a "s oc i al " or a non- soci al '1 nature, with the presumed attendant

    consequence that opinions about t hem can be t est ed r espect i vel y by "s oc i al "

    or by "non- soci al " mea ns. It can i nst ead be def i ned in t erms of t he awar e

    ness (or the degree of subjective probability) that there exist alternatives to

    t he j udgement one is maki ng. A l ow (or nil ) pr obabi l i t y t hat al t er nati ves, t o

    - 7 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    8/64

    one's opi ni ons exi st ma^ be due t o t he consi st ency over t i me in t he checki ng

    of t hese opinions t hrough non- social moans , as in Fest inger 's exampl e of

    fr agil it y and hammer ; but it may al so be due t o t he very hi gh soci al cons ensus

    about t he nat ur e of a phenomeno n, i ndependent l y of whet her t he phenomenon

    is thought of as being "phy s i ca l ", "nat ur al 11 or "s oc i al ". It is undoubt edl y

    tr ue that cert aint y c an very often be more easi l y r eached about t he phys i cal

    t han about t he s ocial means of t est i ng, but t his is not a t heor et ical dist incti on

    bet ween what appears and does not appear as 'obj ect iv e r eal i t y". It cannot

    be s ai d t hat a human or gani sm t urns t owar ds soci al means of val i dat i ng

    opi ni ons onl y when non- soci al means for doing so are not avai l abl e . There

    ar e many e x ampl es , both in t he his t ory of s ci ence in our own cul t ure and int he s ys t ems of knowl edge of other cult ur es, of procedur es whi ch fol l ow t he

    opposit e cour se; i . e ; . t hey do not use t he means of "phys i cal " t est ing whic h

    ar e, in pri nci pl e, avail abl e becaus e of t he very hi gh (or compl ete) soci al

    consensus about the nature of a phenomenon.

    Therefore, "social comparison processes" have an even wider range of

    appl icat ion t hanFest inger was wil l ing to ass i gn to t hem. The range of appl i

    cati on i ncl udes both t he s ocial cont ext (or si gnif i cance) of "non- soci al "

    t est i ng, and t he cas es wher e the hi gh social consens us about t he nature of a

    phenomenon i s suff i cient t o confer t he mark of "objec t i vi t y" on opi ni ons about

    it . In hi s t heor y, Fest i nger was mainl y concer ned wi t h t he soci al t est ing of;

    opi ni ons about charact eri st i cs of i ndiv i dual s , and wi t h t he r esul t ing "r el ati ve

    simil arit y i n opi ni ons and abil it i es among pers ons who ass oci at e wi t h one

    another (at least on those opinions and abilities which are relevant to that

    associ at i on) ". The theory was primaril y addres sed at the wit hin- group eff ects

    of t he pr ocess of soci al compar i son (such as pres sures t owards unif ormit y i n

    a group) whi l e "compar i sons wi t h members of a dif fer ent st atus gr oup, eit her

    higher or l ower may somet i mes be made on a phant asy l evfel, but v ery r arel y

    in r eal i t y". Though Fest i nger qual if ies t his st atement by addi ng that compari

    sons bet ween groups t hat dif fer are not compl et el y el i mi nat ed, t he focus of

    hi s di sc uss i on r emai ns on i ndi vi dual s compari ng t hemsel ves wit h other

    in div idu als .

    On the basis of our discussion so far, we are now able to make two

    l b i l i i i i l i i

    - 8 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    9/64

    function "as a system of orientation which creates and defines the individual's

    own pl ace in so ci et y". The fir st concer ns the "obj ect i ve r eali t y" of compari

    sons focusing on an individual as an individual and comparisons based on an

    i ndiv i dual 's member ship of a part icul ar social gr oup. With regard to the first

    i s s ue, it can be s aid that t he onl y "r eal i t y" t est s t hat matt er wi t h r egar d to

    group characteri st ics are tests of social reali t y . The characteri st ics of one's

    group as a whole (such as its st at us, its r i chness or povert y, i t s ski n col our

    or its ability to reach its aims) achieve most of their significance in relation

    to per cei ved dif f er ences fr om other gr oups and t he val ue connot at i on of t hese

    dif f er ences. For ex ampl e, economi c depri vati on acqui r es its i mport ance in

    soci al at t i t udes, i nt ent ions and act i ons mainly when it becomes 'rel ati ve

    depri vat i on"; eas y or dif fi cul t ac ce s s to means of product i on and consumpt i on

    of go ods , t o benefi t s and opport unit ies becjpme psychol ogi cal l y sal ient

    mainl y i n r el ati on t o compari sons wi t h ot her gr oups; t he defi ni t i on of a group

    (nat i onal , r acial or any other) makes no sense unl ess t here are other gr oups

    ar ound. A group beco mes a gr oup in t he se nse of being per cei ved as havi ng

    common characteristics or a common fate only because other groups are

    present in t he envir onment .Thus, t he psychol ogi cal aspec t s and consequences of the member shi p of

    a group are capabl e of any ki nd of a defi nit ion onl y bec aus e of their i nsert ion

    int o a mul t i- group str ucture. Consequent l y , t he social ident it y of an indi vi dua

    conceiv ed as his "knowl edge that he bel ongs t o cert ain social gr oups toget her

    wit h s ome emot i onal and v al ue s i gni f i cance t o hi m of hi s member s hi p can

    only be defined through the effects of social categorizations segmenting an

    i ndi vi dual 's soc i al envi r onment i nt o hi s own group and ot hers . A soci al group

    wi l l , t her ef or e, be c apabl e of pr es er v i ng i t s cont r i but i on t o t hos e as pe c t s of

    an individual's social identity which are positively valued by him only if it

    manages to keep it s posi t iv el y val ued dist i nct i veness from other gr oups. It

    is t r ue, of cour s e, that somet i mes an indivi dual may wi s h for hi s own group

    t o be more si mil ar t han it is t o cert ain other gr oups; t his i s usual l y s o when

    t hese groups are c onsi dered as ''super i or1' or "bet t er " in some r es pect s.

    Howe v e r , t he fact t hat an indi vi dual may wi s h for his gr oup t o be more l i ke

    another in cert ain r espect s means t hat , i n t hese r es pect s . his own group is

    - 9 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    10/64

    - 10 -

    not adequat el y f ul fi l l ing it s f unct ion of cont ri but ing to posit i vel y val ued

    soci al i dent it y. Ther efor e, if it is tr ue t hat s ocial compar i sons on t he i ndi

    vi dual l evel f ocus , as Fest inger st at ed, on cnming closer t o and ass ociat i ng

    wit h t hos e who ar e si mi l ar t o us , s oc i al compar i s ons be t ween gr oups ar e

    f ocused on t he est abl i shment of dist i nct i veness bet ween one's own and

    other groups.

    The first studies on social categorization by the applicant and his

    col l eagues , bri ef ly summari zed earl ier in this proposal , provi ded - at best -

    suggest i ve e vi dence t hat some pr oces ses wer e at work i n an int ergroup

    situation which, despite their leading to strong intergroup differentiation

    t hr ough t he behavi our of t he s ubj ec t s , coul d not be att ri but ed eit her t o

    previ ous host i l i t y, or to an "obj ect i ve" curr ent confl ict of interest s bet ween

    the gr oups, or to a si mpl e ver si on of the subject s' sel f- int erest . Thes e

    st udi es wer e in no se ns e experi ment a cr uces; but rat her, t hey ser ved as

    cr ut ches for furt her t hinking about t he is sues i nvol ved. In addi t i on, an

    import ant met hodol ogi cal pr obl em had to be deal t wi t h bef ore t he s ubj ect s'

    behaviour could be seen as a consequence of social categorization into

    gr oups rather t han of inter- indivi dual simil arit y whic h wa s , in t hese exper iments , ass oci at ed wit h this cat egor izati on; i .e . t he crit erion for group

    members hip was a si mil ari t y bet ween the subject s i n t heir per f or mance in

    t he first part of t he experi ment s. There i s a good deal of evi dence (cf.

    Byr ne, 1971, f or a summary) t hat int er- indivi dual si mi l ari t i es, ev en when t hey

    are fairly trivial, do lead the subjects in constricted experimental situations

    t o " p r ef er t h os e wh o are more "l i ke" t hem.

    For t his r ea s on, furt her experi ment s wer e conduct ed in order to att empt

    a separat i on bet ween t he vari abl es of int er- indivi dual simil arit y and of

    "pur e cat egori zat i on int o dichot omous gr oups. In a t wo- by- t wo des i gn,

    Bil l ig (1972; al s o cf . Bill ig & Taj f el , 1973) c ompar ed his subj ect s' behavi our

    towards others when they were explicitly divided into groups with their

    behavi our when divi si on int o groups was not made expli ci t . He adapt ed for

    hi s pur pose t he pr ocedur e pr evi ousl y use d in t he experi ment s by Taj fel et al .

    (1971). In t he fir st part of Bi l l i g's exper i ment , t he subj ect s wer e as ke d to

    expr es s t heir pr ef er ences for one or t he other of t wo painter s (Idee and

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    11/64

    Kandi nsky ) on t he basi s of a number of reproduct i ons of paint i ngs whi ch were

    shown to them on a screen. In t he s econd par t , in condit i ons whi ch insured

    anonymit y of previ ousl y ex pres sed pr ef er ences, each subj ect was ask ed to

    awar d poi nt s (whi ch had monet ar y val ue) be t wee n t wo anonymous other

    subj ect s who wer e desi gnat ed by code numbers ; There were four experimental

    condi t i ons. In one condit ion (categori zat ion and simil arit y - CS) each subj ct

    awar ded poi nt s t o t wo ot hers , one of whom was - n his own gr oup, t he group

    member ship being bas ed on t he pr evi ousl y expr ess ed pr efer ences (the "Kl ee

    gr oup" or t he "Kandins k y gr oup") , and one in t he ot her gr oup. In t he s ec ond

    condi t ion (categor i zat ion wit hout simil arit y - C S ) , the subj ect s awar ded point s

    t o t wo others who wer e al s o ass i gned to two gr oups (one of whic h was t he

    subj ec t s own group) but t hi s ass i gnment was expl i ci t l y made r andom, and

    had nothing to do wi t h pr evi ousl y ex pres sed pict ure pref er ences. In t he thirdy

    condi t i on (simi l ari t y wi t hout cat egori zati on * C S ) , t he s ubj ect s awar ded points

    to two others whose code numbers indicated that they preferred one or the

    ot her pai nte r , but t he noti on of "gr oups" was not i nt r oduced or ment i oned at

    any point dur i ng t he experi ment . In t he fourt h condi t i on (no cat egor i zati onA'.

    and no simil ari t y - CS) points wer e awar ded t o t wo other s ubj ect s wi t hout

    any ref er ence eit her t o group members hip or to pict ur e pre f er ences. The

    resul t s wer e as f ol l ows: in condi t ions of CS and CS a s ignif i cant amount of

    favouritism was shown towards others who were in the same group as the

    subj ect making t he awar ds; in condit ion CS , t here was some t endency by the

    subj ect s t o favour t hose whos e pref er ences wer e simil ar to their own, but

    t his t endency di d not r each t he l evel of stat ist ical si gnif i cance; in condi t ion

    CS there was no bias in favour of one or the other of the recipients of the

    awar ds. The highly signif icant r esul t s i n condit ions CS repl icat ed t hose

    obt ai ned in t he init ial experi ment s (Tajf el et a l . , 1971). But our mai n i nt erestr/ /*/

    her e i s in t he compari son bet ween condit i ons CS and CS . The favour i t i sm

    shown t owar ds t hose who were as s i gned t o t he subj ect's own group wi t hout

    any r efer ence t o simil arit y in pr efer ences (the CS condit ion) was consi der abl y

    and si gni f i cant l y st ronger t han t he non- si gni f i cant t endency shown i n t he CS

    condition to favour those who, without any reference to their categorization

    int o gr oups, wer e simil ar to t he subj ect in t heir pr ef er ences .

    - 11 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    12/64

    a/

    It cannot be s ai d, of cour s e, t hat t he subj ect s i n condi t i on CS di d not

    cat egori ze as "gr oups " on t he basi s of simil ari t ies of pref er ences. But the

    point of the experiment was that this was not an explicit categorization.

    Thus, the introduction of an explicit social categorization in condition CS -

    whi c h wa s not ba s e d on any pr ev i ous si mi l ar i t i es be t ween t he i ndi v i dual s

    i nvol ved - was much more eff ect i ve in produci ng f avour i t i sm t han t he intro-

    duct i on i n condi t i on CS of simil arit y bet ween indi vi dual s whi ch was not

    rel ated to an expl ici t social categori zat i on. Thes e concl usi ons wer e

    conf i r med i n a st udy by Chas e (1971; al s o cf . Hor nst ei n, 1972) who empl oyed

    a modi fi cati on of t he pr ocedur es use d in t he init ial experi ment s (Taj f el ,

    1970) wi t h gr oups of subj ect s i n Ne w York. As i n Bi l l i g's ex per i ment s, no

    expl ici t categor izat i on int o gr oups was intr oduced; consequent l y, l it t l e or

    no discrimination was found.

    Before going further, it may be worthwhile to emphasize the crucial

    dif f erences bet ween the set s of resul t s we obt ained and t hose obt ained in

    previous work which is the nearest in its conceptions and methods to the

    st udi es descr i bed here: Sher i f 's wor k on int ergroup confl i ct . Sher i f's aim

    wa s t o i nves t i gat e t he ef f ec t s of an ex pl i ci t l y and cl ear l y i nt r oduc ed zer o-

    sum confl i ct bet wee n groups on out group at t i t udes and the s ubsequent

    behavi our of hi s subjec t s. In addi t i on, i ngroup affi l iat ion and out group

    hosti l it y wor e bot h i nt ensi f i ed t hr ough prol onged i nt ragroup int eract i on

    bet ween t he subj ect s. In our experi ments , there was no ext ernal l y def i ned

    confl ict ; if t here was competi t ion (i .e . act i ons aimi ng t o dif fer ent iat e

    bet ween the gr oups i n f avour of one's own) , it was full y and act i vel y br ought

    int o t he sit ua t ion by the s ubj ect s t hems el v es , as soon as t he noti on of

    "gr oup" was i nt r oduced by t he exper i ment er s. The subj ect s wer e never

    t oget her as a "gr oup"; t hey neit her inter acted nor did t hey k now who was in

    t heir o wn gr oup and who i n t he other; t here we re no expl i ci t soci al pres sur es

    on t hem t o act i n f avour of t heir own gr oup; and in no way was t heir own

    i ndi vi dual i nt eres t engag ed i n awar di ng more money to a member of their own

    gr oup. On t he cont r ary , a consi st ent us e of t he maxi mum j oint profi t s t rat egy

    woul d hav e l ed t o al l of t hem r ecei v i ng mor e money from t he exper i ment er s .

    - 12 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    13/64

    It i s t he as s umed need for dif fer ent i ati on (or t he est abl i shment of

    psychol ogi cal dist i nct i veness bet ween t he groups) whi ch s eems t o me to

    pr ovi de, under some condi t i ons, t he major outcome of t he s equence s ocial

    categori zat ion - social ident it y - social c ompar i son. Rel at ed phenomena

    can be shown to exi st in a large vari ety of social si t uat i ons. One major

    exampl e i s pr ovi ded by t he wi der soci al cont ext s i n whi ch the noti on of

    "r ace " i s used as a crit erion for soci al cat egori zati on. For a number of

    r easons "r ac e" has become a val ue- l oaded t er m, a noti on whic h has "sur pl u

    val ue connot at i ons. It may t herefore be i nst ructi ve to ident if y the soci al

    situations in which this notion tends to be used, or "the kinds of social

    differentiations in which subjective social distinctions have been made"

    | l ex, 1969). Accor di ng to Rex , t hese are as f oll ows:

    "1. The si t uati on of cul t ure cont act bet ween peopl es wi t h an adv anced

    indust r ial and mil it ary te chnol ogy, and hunt er s, past oral is t s and agricult ur

    al is t s at low er lev els of dev elop men t .

    2. The sit uati on on a sl ave pl ant ati on.

    3. Cl as s sit uati ons in the cl ass i c Mar xi st or Weber i an scene in which

    men wi t hi n t he same soci et y have dif ferent degrees of market power .

    4. Stat us si t uati ons i n whic h t here i s a concept of higher and l ower

    but in whi ch men are t hought of as fal l ing s omewher e on a cri t erion of est eem

    rat her than in mut uall y ex cl usi ve gr oups.

    5. Sit uati ons of ethnic pl ur al i sm in whi ch gr oups wit h dif feri ng cul t ures

    and/or phys i cal charact eri st i cs work t ogether i n t he same economy but r etain

    their social and cultural identity.

    6 . Sit uat i ons in whi ch a minor it y group occ upi es a par i ah or scape- goat

    r ol e." (Rex, 1969, p. 147).

    In t hree of t hese si x si t uat i ons, val ue dif fer ent iat ions bet ween groups

    or indi vi dual s are expl icit l y stated ("l ower l evel s of deve l opment ", "crit erio

    of es t ee m", "par i ah"). In t he r emaining t hree, t hey are not far bel ow t he

    sur f ace. What ev er i ts other uses may be , t he noti on of "r ace " has become

    in its general social usage a shorthand expression which helps to create,

    r efl ect , enhance and perpetuat e the percei ved di ff er ences i n "wor t h" bet ween

    human gr oups or i ndi vi dual s . It cont r i but es t o maki ng thes e dif f er ences as

    - 13 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    14/64

    c lear- c ut and inf l ex ibl e as p os s i bl e . Therefore, its application in the wide

    r ange of s oci al contexts enumerat ed by Rex wi t nes se s to the i nt r oduct i on,

    whene v er po s s i bl e , of di f f er ent i at i ons i n t er ms of v al ue whi c h i nc r eas e t he

    distinctiveness of social categories and thus contribute to their function as

    a guide for social act ion.

    This establishment of dist inctiveness is by no means, however, confined

    to sit uati ons connect ed wit h t he noti on of r ace . It f i nds it s way , for ex ampl

    int o t he c ompl ex eff ect s t hat cul tural and soci al r el ati ons have on t he mut ual

    comprehensi on and acc ept ance by inter acti ng gr oups of t heir l anguages and

    dial ect s . Fis hman (1968) recent l y wr ot e, basi ng his st atement on l ingui st i c

    evi dence f rom We s t Afr i ca, t he Swahil i r egion of Cent r al and East Afr ica ,New Gui nea, Scandi navi a and Sout h East Asi a: "Di vi s i ve ness is an i deol

    ogical posi t i on and it can magni f y minor dif f er ences; i ndee d, it can manu

    fact ure dif fer ences i n l anguage as in other matt ers almost as easi l y as it

    can capit al i ze on more obvi oj s di f f er ences. Simil arl y , unif icat ion is al so

    an ideol ogi zed posit i on and it can mi ni mi ze s eemi ngl y major dif f er ences or

    ignore them ent i r el y, whet her t hese be in t he r eal m of l anguag es , r el i gi on,

    cul t ur e, r ac e, or any other basi s of dif fer ent iat ion" (p .45). Fis hman's

    "i deolog i zed posi t i ons" are posit ions i n whic h simil arit ies or di f f er ences,

    whi c h coul d i n pr i nci pl e be ent i r el y "neut r al 11 (e .g . bet w een lan g u ag es ,

    l andsca pes , f l ags , ant hems , post age st amps, footbal l t eams and al most

    anything else) become endowed with emotional significance because they

    rel ate to a superordinate v al ue, such as is t he cas e wit h nati onal i sm in

    Fishman's own discussion.

    The present research proposal postulates that the reason for this

    behavioural and evaluative intergroup differentiation is in the need that the

    subjects have to provide order, meaning and social identity to the experi

    ment al si t uati on; and t hat t his need is fulf il l ed t hrough t he creati on of

    int ergroup dif fer ences when s uch di f fer ences do not in fact ex i st , or the

    att ri buti on of val ue t o, and t he enhancement of , what ever di f fer ences t hat

    do exisfft.

    - 14 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    15/64

    I

    II. Int argr oup hypothes es In socl rJ nont exte

    The sequence of social categorization - social identity - social

    compari son - psychol ogical di st i nct i veness wil l now be di sc uss ed from the

    point of v i ew of hypot heses whi ch can be der i ved fr om i t . Thi s wi l l be done

    in t wo st ages: f i rs t , in terms of examples i n rel at ivel y concret e social

    cont ext s whi ch s eem t o provi de intuit ive s uppor. t o t he causal se quence

    out l ined above; and se cond, in t erms of r esear ch hypot heses whic h wil l

    guide the st udi es ar i si ng from t he pr esent proposal .

    Social categor i zat i on wil l be consi der ed here as t he first l i nk in t he

    sequence (although some social situations create a need to categorize arising

    out of a variet y of prel i minary condit i ons, in whic h cas e the act of cat egor

    i zing the soci al worl d int o gr oups becomes an eff ect of t hese condi t i ons).

    In our experiments which were previously described, as in those of Sherif

    and in other r el ated st udi es , t his categori zat ion was provi ded by t he exper i

    menter s. If t he exper i menter s' arbit rary intr oduct i on of dichot omous cat e

    gori es l eads to dif ferent ial int ergroup behavi our fol l owi ng t he li nes of t he

    i mposed di vi s i ons, one c an concl ude that t here is an appr oximat e i dent it y oft he experi ment er s' and t he s ubj ect s ' st ruct uri ng of t he soci al worl d i n t erms

    of dif ferent gr oups. Thi s i s not a tautol ogy, si nce there is no a pri ori r eason

    why t he s ubj ect s s houl d be ha v e i n t er ms of a di vi s i on i nt o gr oups ba s ed on

    gues s i ng numbers of dot s , prefer ri ng one or anot her painter or - a s i n t he

    extr eme ca s e of Bi l l i g's (1972) exper i ment s - on t he t oss of a coi n.

    When one l eav es t he "mi nimal " si t uat i ons, t he pr obl am of def i ni ng or

    ident if ying the pre- exist ing categori es becomes much easi er , s i nce the

    informati on about t hem can be obt ained eit her dir ectl y f rom t he s ubj ect s or

    by observing "natural" situations, or - ideally - by coordinating both these

    sources of information.

    Our interest in the concept of social identity, as defined earlier in

    t he pr oposal , i s not in att empt s to descr i be it for "what it i s " in a st atic

    sense - a daunting task which has baffled many social scientists of various

    per s uasi ons and for whi ch one l acks both opt i mi sm and t emer i t y. Social

    i dent it y is under st ood her e as an inter veni ng ca usal mechani sm i n si t uat ions

    - 15 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    16/64

    of social change (cf. Taj f el , 197 2b) - obser ve d, anti ci pated, f ear ed, des i r ed, or

    pr epared by the i ndi vi dual s i nvol ved; and t he effect s of these c hanges on their

    s ubse quent i nt ergroup behavi our and at t i t udes. From t hi s point of v i e w, t hree

    categories of situations appear crucial:

    (i) The badl y defi ned or margi nal soci al si t uati on of a gr oup, whi ch

    pr esent s t he indi vi dual s i nvol ved wi t h di ff icult ies of defi ni ng their pl ace i n a

    social system;

    (ii) The gr oups social l y defi ned and consensual l y accept ed as "super i

    at a point of time when this definition is threatened either by occurring or

    i mpendi ng social chang e, or by a confl ict of val ues i nherent in t he "super i or i t y";

    (iii) The gr oups social l y def i ned and consensual l y acce pt ed as inferi oat a point of t i me when - for what ev er r eas ons - eit her (a) member s of a gr oup

    have engag ed in a s har ed pri se do consc i ence of t heir inferior s t atus; or (b)

    they have become aware of the feasibility of working towards alternatives to the

    exi st i ng s i t uat i on; or a combinat i on of (a) and (b) , whi ch may al s o imply (a)

    l eadi ng to (b) , or (b) l eadi ng t o (a) .

    The "dynami c" appr oach to problems of soci al ident it y adopt ed in t his

    dis cuss i on is based on several consi der ati ons. First , i t is unl ikel y that there

    exist many examples of intergroup situations which are static in the sense that

    t hey consi st of an unchangi ng set of soci al r el ati onships bet ween t he groups.

    We ar e, ho wev er , l es s concer ned her e wit h s oc i al s i t uat i ons t han wit h t hei r

    psychol ogi cal count erpar t s; t hese are bound t o be even l ess st at i c. This

    becomes quite clear when one reconsiders briefly for the purpose at hand the

    focal problem of t his proposal : that of soci al i dent it y under st ood as der i vi ng

    in a compar at i ve and "r el at i onal " manner fr om an i ndiv i dual 's gr oup member shi ps.

    For t he pur pose of our ar gument , one c an dis t i nguis h bet ween "s ec ur e"

    and "i nsec ur e" soci al i dent it y. A compl et el y secur e social ident it y woul d impl y

    a rel at i onship bet ween t wo (or more) gr oups i n whi ch a change in the t exture of

    psychol ogic al di st i nct iv eness bet ween t hem is not concei vabl e. For an

    "i nf er i or" group t hi s woul d impl y the exi st ence of a total c onse nsus about t he

    nature and t he fut ur e of their inferiori t y; in other wor ds , to fet ur n t o our previ ous

    dis c u s s ion (c f . p . 8 of "soci al r eali t y" as rel ated to Fest inger's theory of

    soci al compari son (1954) , ther e woul d have t o exi st a complet e psy chol ogi cal

    - 16 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    17/64

    "obj ect i f i cat i on" of a s ocial st atus quo wit h no cognit iv e al t ernat i ves of any

    kind avail abl e to chal l enge the exist i ng soci al real it y . It is poss i ble that

    historians and social anthropologists could provide some relevant examples in

    compl etel y st abl e and is olat ed societ ies ; t hese examples coul d har dl y, howeve r

    f i nd t heir count erpart i n most of t he cont empor ar y wor l d.

    A compl et el y s ecur e s oc i al i dent i t y for a gr oup c ons ens ual l y super i or i s

    almost an empir ical i mpossi bi l i t y . The kind of psychol ogical di st i nct i veness

    t hat insures t heir unchal l enged superi ori t y must not onl y be gai ned; it must al so

    be pres er ved. And it can be pr eser ved onl y if social condi t i ons of dist i nct i venes

    are carefull y per pet uated, toget her wi t h t he si gns and sy mbol s of di st i nct i ve

    status without which the attitudes of complete consensus about superior distinc

    t i veness are in t he danger of dis integr ati ng. In t his s ens e, t herefor e, even in

    the most rigid oasto system (be it racial or any other) , the social distinctions

    whi c h may appear ver y s t abl e ar e r el at ed t o a cont i nuous l y dy nami c ps y chol og i c a

    situation in which a superior group can never stop working at the preservation

    of its distinctivenoss. It is very difficult to think of cases of intergroup relation

    whi c h woul d pr es ent ex cept i ons t o t his s t at ement , apar t per haps fr om i nf ant s

    and t eacher s i n a nurs ery sc hool. (Thi s al so happens t o be t he excepti on for

    whi c h it i s di f f i cul t t o i magi ne t he pos s i bi l i t y of a s us t ai ned and s oc i al l y s har ed

    pattern of intergroup discrimination based on hostility).

    A mor e s er i ous ex ampl e can pr oba bl y be f ound i n t he not i ons abo ut t he

    'nature" and the relative roles and positions of men and women prevailing in

    some cul t ures and some hist orical per i ods. In t hese c as es , t he mass i ve

    accept ance by both s i des of cert ain kinds of psychol ogical intergroup dis t i nct iv e

    ness pr event s t he occur r ence of ser i ous and soci al l y shar ed ident it y probl ems.

    I t is , h ow ev er , interest ing to se e that as soon as t hese accept ed noti ons are

    ser iousl y chal l enged, t he inter group att it udes undergo cert ain changes whic h are

    in l ine wit h the pr esent ar gument . The psychol ogical dif ferent iat i ons of s ex es ,

    as l ong as t hey are hi ghl y c onse nsual , ar e not accompani ed by , or r el ated t o,

    att it udes of int ergroup hosti l it y; howev er , t he new search by an act i ve minori ty

    in t he "i nfer i or" group of dist i nct i veness on an equal l evel cr eat es, in some

    cases, explicit outgroup hosti l ity on one side and equally hosti le defensive

    r eact i ons on t he other (Doi s e & Wei nber ge r , 1973). The i mpact of t he i mpl i citand socially shared problems of identity which are involved can perhaps be

    - 17 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    18/64

    gauged f rom t he ext ent of cover age gi ven by var i ous communi cat i on medi a t o all

    ki nds of sci ent if ic and pseudo- sci ent i f i c pr onouncement s about t he "nat ure" of

    p s y c h olog ic al s ex di f f eren c es .

    For all the reasons outlined above, this research proposal will thus be

    concerned wit h cas es of insecure social ident it y. These wil l be di sc uss ed in

    t erms of a t wo- by- t wo t abl e, in whic h t he t wo crit eria for cat egor i zat i on are:

    consensual l y "super i or " ver sus "inferi or" gr oups; and t he i ndiv i dual 's abil it y

    t o "pa s s " fr om one group t o anot her .

    The f irst of t hese criteria is j ust if ied bec aus e, as wil l be seen l ater ,

    di ff erent hypot heses pert ain t o t he t wo ki nds of gr oups . The s econd cri t erion

    appear s bas i c on a_ pri ori g r ounds. In each indi vi dual 's l if e t here wi l l be si t u

    ati ons in whic h he act s , excl usi vel y or mai nl y, as an i ndi vidual rather t han as

    a member of a group; t here wil l be others i n whi ch he ac t s , exc l usi vel y or

    mai nl y , i n t er ms of his gr oup member s hi p. One of t he i mport ant deter mi nant s

    of an indi vi dual 's choic e to act in terms of sel f rather t han in t erms of hi s gr oup

    is what we shall refer to in this discussion as "social mobility" as contrasted

    wi t h "soc i al c ha ng e ". The f or mer r ef er s t o s i t uat i ons i n whi c h i t i s r el at i vel y

    eas y t o move i ndi vi dual l y fr om one s oci al gr oup to anot her ; so t hat , i f a groupdoes not contr ibute adequat el y t o an i ndiv i dual 's soci al i dent i t y, one of the more

    obvious solutions for him is to move, or attempt to move, to another group.

    In the latt er cl ass ar e t hose s it uati ons in whi ch, for what ever r eas ons, "pass i ng

    fr om one group t o anot her i s ver y dif fi cul t or imposs i bl e. It may bo expect ed

    t hat , i n t hese si t uat i ons, t here wi l l be many occa si ons ( and const raint s)

    l eading an i ndivi dual t o act a s a member of hi s g r oup, or at l east in t he knowl edg

    that he is cat egori zed as s uch. Social change (as dis t inct from soci al mobil it y)

    ref ers t herefore in this di sc uss i on to changes in t he r el at ionships bet ween the

    groups as a whol e, to expect at i ons, fears or desi r es of such changes , t o act i ons

    ai mi ng at i nduci ng or pr event i ng t hem, or t o i nt ent i ons and pl ans t o engage i n

    t hese act i ons. The psychologi cal counterpart of social change, in the l imit ed

    s ense of t he term adopt ed her e, is t herefore i n t he indi vi dual 's awar ene s s that

    many import ant as pect s of his l i f e, i ncl udi ng t he acqui si t i on or pr eserv ati on of

    an accept abl e s ocial i dent i t y, can onl y be bas ed on a change (or r esi st ance to

    change) in t he i mag e, posi t i on or ci r cumst ances of hi s gr oup as a whol e.

    - 18 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    19/64

    I

    The result ing t wo- by- t wo cl assi f i cat i on of cas es and predict ions ari si ng

    fr om t hem can t heref ore be pres ente d as f ol l ows:

    Insecure Intergroup social comparisons

    Condi t i ons conduciv e Condit i ons conduci ve

    t o l eavi ng one's group t o st ayi ng i n one's group

    C on s en s u al ly

    superior groups

    - 19 -

    C on s en s u al ly

    inferior groups

    1) S uper i or gr oups (Boxe s A and B)

    Insecure social compar i sons ar is i ng wit hin a group whic h is consensual l y

    defined as being of higher status can bo due to two sets of conditions:

    (a) The gr oup's superi or stat us i s threat ened (or per cei ved as t hr eat ened)

    by another group;

    (b) The superior st atus is rel ated to a confl ict of val ues: i .e . i t is

    concei ved by some as base d on unfair adv ant age s , var ious other forms of

    i nj us t ic e, expl oit at i on, i l l egi t imat e us e of f or c e, et c .

    In t he cas e (a ), Box A is not l ikel y to cont ain many i nst ances a s l ong as

    t he threat does not become ov er whel mi ng. It wil l cont ain hardl y any i nst ances

    in si t uati ons in which "pas s i ng" is very dif f i cul t (e .g . an apart heid societ y).

    In Box B one c an pr edi ct an i nt ensi f i cat i on of precaut i ons ai mi ng t o keep t he

    superi or group in it s posi t i on. On t he l evel more specif i cal l y rel ated to t he

    present discussion, the level of social intergroup comparison, one can predict

    t he cr eati on and use of new condit i ons whi ch wi l l enabl e the superi or group to

    pr eserve and enhance its psychol ogic al dis t i nct i veness. Thi s may t ake many

    forms, such as social and psychological separation of many kinds, creation of

    a var i et y of dis t in ct iv e s y mbol s , et c .

    In t he cas e (b) , confl ict of v al ues , t wo sub- case s must be dis t i nguis hed:

    (i) The confl ict of val ues is of such intensi t y that it dest r oys t he pos

    contribution to social identity that the group provides. Thi s is Box A whic h

    cont ains "condit i ons conduci ve to l eaving one's own gr oup": e .g . upper or

    iddl l l i i " d " f l l ki d Th il l b

    - - - - - - - - ' ' 1

    A

    - ....... . .. . . . ... "

    B

    c D

    .. ___________________

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    20/64

    her e no discr i mi nat i on agai nst t he out group and no hosti l it y agai nst i t . But t his

    is hardly an interesting intergroup prediction since group membership is often

    cancelled for al l practical purposes, or even sometimes posit ively reversed.

    In t he ca s e of r ever sal , t he "r enega des " become some of t he most act i ve member s

    of the inferi or group and pre sent , oft en in an int ensi f ied f orm, t he charact eri st ics

    whi c h wi l l be di s c us s ed l at er i n r el at i on t o Box D.

    (ii) The confl ict of val ues ex i s t s , but ingraui affi l iat ion is suff ici en

    power f ul to r emai n t he det ermi nant of att i t udes and behavi our. This i s agai n

    Box B. The confl ict of val ues can onl y be r esol ved through fi nding new j ust i

    f i cat i ons for t he mai nte nance of stat us quo. This i s the condi t i on for t he

    cr eati on and adopti on at l arge by the members of a gr oup of new "i deol og i es "

    (e .g . t he "whit e man's bur den", t he "i nherent s uperi ori t y" due to unbri dgeable

    innat e di f f eren c es , t h e "s av i ng o f s oul s ', et c . ) . T hes e ideolog ies repres ent

    t he cr eat ion of new forms of psy chol ogi cal dist i nct i veness and t he enhancement

    of t hose amongs t t he ol d ones whic h are sti ll ser vi ceabl e. The clearer ar e t he

    "obje ct i ve" condi t i ons pr event ing the l eavi ng of one's gr oup (such as raci al

    dif f er ences, powerf ul l y sanct i oned rel igi ous di f f er ences, e t c .) , the more l ikely

    it is that t he confl ict of val ues wil l r esul t i n t he creati on and wi de and easydif fusion of t hese new and enhanced forms of psychol ogical di st i nct i veness.

    A not e must be i nser t ed her e s o t hat t he anal y t i c di s t i nct i ons j us t made do

    not distort the psychological realities of the situations which are being con

    si dered. The act i ons and att it udes di rect ed t owar ds the pr eser vat ion of t he

    st atus quo in t he cas e (a) di sc uss ed abov e ( i .e . t hreat f rom another group)

    cont ain somet i mes t he seeds of a "sec ondar y" confl ict of val ues: an i nt ensi f i

    cation of discriminatory behaviour and of hostile attitudes following upon this

    threat may cl ash wi t h general l y accept ed val ues represent i ng the ''off ici al "

    i deol ogy of a societ y such a s , for exampl e, Chr is t ian et hics . In such ca s es ,

    t here i s a mergi ng of t he eff ect s descr i bed above for ca s es (a) and (b) (i i ). The

    psychol ogical ori gins of t he sit uations ar e, however , dif f erent . In t he case (a)

    t he inferi or group is percei ved as a t hreat to t he st atus qu o . both s ocial and

    psychol ogi cal , and the dri ve towards i ncr eased dis t i nct iv eness aims both at the

    preservation of the previously existing differentiations and (in some cultural

    contexts) at the just if icat ion of consequent act ions. In the case (b) (ii) there

    - 20 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    21/64

    as f or exampl e, i n t he earl y colonial sit uati ons - t he wi de dif f er ences bet ween

    t he gr oups are ther e for al l to s ee . There may even occur an ideali zati on of t he

    out group (cf . e . g . Kier nan, 197 2, about t he myt h of t he "nobl e s av ag e" in t he

    XVII It h cent ur y and l at er) , a s l ong as t he myt h- mak er s and t hos e who ac c ept

    their images have nothing directly to do with the outlandish creatures so

    beaut if ul l y port r ayed. When, however, direct interaction of certain kinds

    beg i ns, in whic h what is done to s ome peopl e i s not usual l y done to others ,

    j us t i f i cat i ons and r eas ons mus t be f ound. Thi s i s , of c our s e, a wel l - known

    and wel l - document ed se quence of event s . Its rel ati on to t he main argument of

    t his proposal i s t hat , as in t he case (a ), the psychol ogical "st r uct ur e" which

    is the most convenient, both cognit ively and affect ively, for resolving the

    ensuing confl ict of val ues and thus preservi ng int act a posit iv e social ident it y,

    i s t he cr eati on and enhancement of t he appr opr i ate f orms of psy chol ogi cal

    di st i nct iv eness bet ween t he gr oups.

    2) Infer ior gr oups ( Boxes C and D)

    (a) Box C: Condit i ons conduci ve t o l eavi ng one's own gr oup.

    Thes e are t he sit uati ons of soci al mobil it y as def i ned earl ier: Ther e is

    enough soci al f l exibil it y to enabl e an i ndi vi dual t o mov e, or hope to mov e, from

    one group to anot her ; t here ar e no seri ous soci al sanct i ons fr om eit her of t he

    gr oups for movi ng; and no ser i ous confl i ct of val ues i nvol ved in movi ng. One

    should expect that after having joined the superior group or even before, some

    i ndi vi dual s wil l work harder t han most at est abl is hing t heir c lear- cut dis t inct

    i veness from t he percei ved i nadequaci es of their past social i dent it y.

    (b) Box D: Condi t i ons conduci ve to st aying in one's own gr oup.

    This box pres ent s a much greater i nt erest from t he point of v i ew of int er

    group att i t udes and behavi our t han t he prev i ous one. The major s oci al condit i ons

    are: any form of cast e sy st em (whet her det er mi ned by bi r t h, r ace or ot her

    cri ter ia); or any other social dif fer ent iat ion sys t em whi ch, for whatev er r eas ons,

    makes it dif fi cul t to move . The t wo major psychol ogi cal condi t i ons are; a

    st rong confl i ct of va l ues i nherent in l eavi ng one's gr oup; or t he fear of power f ul

    soci al sanct i ons for so doing; or both in combinat i on. In most si t uat i ons the

    social and psychologi cal condit ions wil l , of cours e, interact and reinforce

    each other.

    - 21 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    22/64

    The ass umpt i on i s made here (se e p. 16 above) t hat , i n many of t hese

    condi t i ons, t he pr obl ems of social ident it y of t he inferi or group woul d not ne c es s

    ari l y expres s t hemsel ves in social behavi our unti l and unl ess t here is some

    awareness that the existing social reality is not the only possible one and that

    alt ernati ves to it are concei vabl e and perhaps at t ainabl e. I f t his awar eness

    ex i s t s , t he probl ems of soci al i dent it y confr ont i ng t he members of inferior..- groups

    can be sol ved i n one of sev er al way s , or a combinat i on of more t han one:

    (i) To bec ome, t hrough act i on and r eint erpretat ion of gr oup charac t er i

    more l i ke t he superi or gr oup. (It wi l l be r emembered t hat, in v i ew of t he di ff i

    culty of "passing" implied here, this cannot become a widespread individual

    soluti on; it wil l have to apply t o t he group as a whol e).(ii) To reint erpret t he exis t i ng inferior charact eri st i cs of t he gr oup, so

    they do not appear as inferior but acquire a positively valued distinctiveness

    fr om t he s uper i or gr oup.

    (iii) To cr eat e, t hrough social acti on and/or di ff usi on of new 'i deol ogi e

    new group characterist ics which have a posit ively valued dist inctiveness from

    t he superi or gr oup. t/J unJJ un) j iy jwuf yrfiy0 .

    A par ent hes i s must be i nser t ed her e bef or e t he s e t hr ee s ol ut i ons ar e br i ef l y

    consi der ed. It concer ns t he meani ng i n t his di sc uss i on of t he t erms "infer ior "

    and "s uper i or ". Thi s di st i nct ion must be unders t ood, of cour se , in t he cont ext

    of its s ocial deri vati on; for exampl e, bl ack ski n is not , out si de of speci f i c

    soci al cont ext s, either an inferi or or a superi or att ri but e. But It becomes one,

    gi ven certain sooi&l - psychol ogical condi t i ons. In pri nci ple, any group charact

    eri st ic coul d become (and most do) val ue- l aden in t his s ens e. One can not e,

    for ex ampl e, t he per si st ence ev en t oday of bl ond hair and blue ey es in a l arge

    proporti on of t he daunt l ess her oes of war comi cs ; or the s i gni f i cance that l ong

    hair has recently acquired in a variety of social contexts - both for those who

    use it for it s dis t i nct i veness and for these v^o use it as an ident if yi ng si gn of

    moral t urpit ude.

    The f irst solution, which is that of cultural, social and psychological

    assi mi l at i on of a group as a whol e, is somet i mes pos s i bl e. One might even

    pr edict t hat , gi ven favour abl e condi t i ons, it coul d become the sol ut ion t o be

    tr ied fi rst . In order, howev er , for a group as a whol e to succe ed in el iminat ing

    both its social and psychological inferiority one process must f irst take place:

    - 22 --

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    23/64

    t he br eaki ng do wn of t he barr i ers pr event i ng t he group from obt ai ni ng i mproved

    ac ce s s to condi t i ons which.i t coul d not previ ousl y obt ain. As soon as t his

    happens, one of two psychological processes will tend to appear: if the group

    remains separate , a general reinterprotation of its distinctive characteristics in

    new and posi t i vel y val ued t erms; or, al t er nat i vel y, t he break down on both si des

    of t he psychol ogi cal barri ers to "pas s i ng'1. The first of t hese mer ges wi t h t he

    soluti on (ii) whi ch wi l l bo di sc uss ed bel ow; t he sec ond may fi nal l y l ead t o t he

    dis appear ance of a group as it mer ges wi t h anot her . The co nsequences of a

    relative decrease in antisemitism in some countries after the last war exemplify,

    in relation to Jews who do not live in Israel , t he funct i oni ng of both t hese

    p r o c e s s e s .

    It may be expec t ed that t he s ol ut i ons (ii) and (fli) ment i oned abov e wi l l

    appear in conjunct i on, and t hat soci al act i on wi l l be an important i ngredient

    of both; but for t he sak e of empir ical di st incti ons t hey wil l be di sc uss ed

    separat el y. It wil l be remember ed t hat both origi nate in si t uati ons wher e, for

    what ev er r e as ons , t he i nfer i or gr oup i s not abl e t o mer ge wit h t he super i or one ,

    ncr can t he i ndiv i dual members of it l eav e t heir own gr oup and joi n anot her.

    Sol ut ion (ii) i mpl i es t hat , wit h t he pr i se de consc i ence of t he il l egit imacyof a pr evi ousl y conse nsual inferiori t y, a new kind of dist i nct i veness must be

    creat ed on t he basi s of some exi st i ng group charact er i st i cs. The cl earest recent

    example of t he whol e proces s can be found i n t he psy chol ogi cal changes t hat

    are t aking pl ace,arx>ngst t he Amer ic an bl ac ks . The very use of the term "bl ac ks "

    in t his t ext , whi ch woul d have had ver y dif ferent connotat i ons onl y a f ew years

    ag o, al r eady t est if ies to t hese changes . The ol d interpretat i ons of dis t i nct iv e

    ness ar e r ej ect ed; t he old charact eri st i cs are bei ng gi ven a new meani ng of

    dif ferent but equal or superi or. Exampl es abound: t he beaut y of bl ac k nes s ,

    t he Afr i can hai r - do, t he Afr ican cul tural past and t r adi t i ons, t he reint erpret ati on

    of Negr o musi c fr om "ent er t ai nment " t o a f orm or art whi ch has deep root s i n a

    separ at e cul tural t r adit ion; t he t aking over or r e- cr eati ng, at one time, of cert ain

    aspect s of ideas about nig r i t ude, et c. At t he same t i me, t he ol d att empt s to be

    "a l it tl e mor e l i ke t he other gr oup are pr oudl y r ej ect ed, no more st r aighteni ng

    of hair for beaut i f ul bl ack gi rl s or usi ng of var i ous pr ocedur es fear l i ght eni ng t he

    sk i n. The ac ce nt s , dial ect s, sway of t he body, rhythms of danci ng, t exture of

    - 23 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    24/64

    the details of interpersonal communication - all this is preserved, enhanced

    and r e- eval uated. The pri se de consci ence st art s, as it is often t he c as e,

    wit h an ac t i ve mi nori t y (c f . Mos c ov i c i 19 73) . As t he new- f ound di s t i nct i v enes s

    does do its job of creating a positive and healing new version of social

    i dent i t y, t he pr edi ct ion can be made that all it s f orms wil l f i nd an eas y and

    wi de s pr ead di f f us i on at l ar ge.

    Solution (iii) , the creation or invention of nev7 characterist ics which

    est abl is h a posit ivel y valued group di st i nct i veness, is structurall y simil ar

    t o sol ut i on (i i ), Exampl es of it can be found i n t he devel opment of new

    nati onal is ms (cf . Taj fe l , 1969b, 1970b). "In many new nations the need is

    felt t o st r ess or cr eate common bonds i n order t o for ce the pac e of t he devel op

    ment of nat i onhood. The forgi ng of bonds need not be of a "r ac i al " k i nd, .

    though it has often been of this nature, particularly in the young European

    nat i onal i sms of t he Xl Xt h cent ur y. The phenomenon is even cl ear er in r aci sm,

    old or new; t he r acist i deol ogi es have al way s been char act eri zed by a fr anti c

    sear ch for common bonds of an "i nnat e or "inst i nct i ve" nature in t he di st ant

    past so as to provide a justification for the claim of the special sort of unity

    that the racial group is supposed to have and of its inherent and immutabledif fer ences from other such gr oups ". (Taj fe l , 1969b, p .l 39 ;c f . Sh af er , 1 9 5 5 ,

    for a gener al dis cuss i on of the creati on of vari ous categor ies of nat ional myths).

    The crefction of new distinctive characteristics implies however a new

    probl em. This probl em al s o exi st s i n some measure i n t he proce ss of the

    r eeval uat i on of t he exist i ng character i st ics (sol ut ion (ii)) , but it becomes

    part icul arl y cl ear when new forms of dist i nct i veness need to be eit her i nvent ed

    or creat ed t hrough act i on. It has been post ul at ed t hr oughout t his dis cuss i on

    that the aim of posit ively valued psychological dist inctiveness is to achieve

    an adequat e form of soci al i dent it y; and t hat the onl y means by whi ch this

    aim can be attained is in the establishment of appropriate kinds of intergroup

    compar i son. There are two stages in t his process whic h, i deall y , bot h need

    to be succes sf ull y r eal i s ed. The fir st (whic h is a condit ion si ne qua non for

    t he s ucces s of t he enter pri se) is t he posi t i ve eval uat i on by the ingroup of its

    newl y cr eat ed charact eri st ics . The second stage consi st s of t he accept ance

    by t he out group of t his eval uat i on. The is s ue i s , however , sl ight l y more

    - 24 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    25/64

    complicated. The new characteristics of the inferior group can be of two kinds:

    (a) They may consi st of att ri but es whi ch are al r eady consensual l y high

    val ued by both (or moisj gr oups , and whi ch t he inferi or gr oup was previ ousl y

    deemed not t o po s s e s s . In t his c as e, t here is no probl em of r e- evaluat ion

    of attributes. The soci al compar i son pr obl em for t he inferi or group i s: wil l

    t he ot hers ac knowl edge t he new i mag e, separ at e but equal or super i or , on

    consensual l y val ued di mensi ons? An exampl e cai. per haps be found in some

    new and wi del y di ff used aspect s of Jewis h ident it y (cf . Her man, 197o). Amongst

    aspect s of group ident it y unaccept abl e to t he young post- war gener at i ons of

    J ews was their el ders ' pass i ve acc eptance of a whol esal e s l aught er of a peopl e.

    The ex cept i ons, such as t he upri si ng of t he ghett o of War s aw or t he revolt s i n

    Tr ebl i nka and other concent r at i on ca mps , beca me t herefor e cr uci al sy mbol s;

    so has t he Mas ada st ar y, l ong back from t he pas t , and in t he pr esent t he

    mil it ary pr owess of t he new st ate of Isr ael . One may per haps be permit t ed to

    i ncl ude an anec dot e, not i n any s ense as any form of proof but a s an il l ust rati on

    of a theoret ical ar gument . Some ti me ago I found mys el f at t he Pl ace Denfert -

    Rocher eau in Pari s in t he mi ddl e of one of the numer ous pol i t i cal demonst r at i ons

    of t he Spri ng 1973. These are occas i ons at whic h ever yone t al ks to ever yonee l s e , and next t o me at t he table of a cafe a heated argument about t he Mi ddl e

    East confl ict devel oped bet ween t hree mi ddl e aged men: t wo It ali an gauchi st e

    intel l ectuals and an Israeli secondary school t eacher . Aft er a ser ies of

    successive explosions on both sides, the Israeli f inally produced his own

    ambi guous and confl i ct ed st atement of new ident it y whi ch went ver bat i m as

    f ol l ows: "Mai nt enant que nous avo ns prouve que nous pouvons otre des

    as s as s i ns comme t ous l es aut r es, tout le monde nous r es pec t e.1' ("Now that

    we ha v e pr ov ed t hat we c an be as s as s i ns l i ke ever y body e l s e , ev er y one

    r espect s u s "). The confl ict i nherent i n t his st atement i s a confl ict of t wo ki nds

    of social ident it y: on t he one hand, the great cul tural and social t radit ions

    of J ewi sh humani s m, se en as di ff erent but equal or superi or; and on t he ot her ,

    the socially comparative justification of new aspects of identity forged in the

    aft ermath of a t r agedy and s een as appr opr i ate i n v i ew of the r ecent pas t or of

    the present polit ical s ituation.

    - 25 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    26/64

    (b) The sec ond cas e concer ns si t uati ons in whic h t he new charact eri st

    of an inferi or group are not consensual l y val ued, t o begin wi t h. The social

    comparison problem of the second stage for the inferior group then becomes:

    wil l t he ot her s ac k no wl edg e t he ne w i ma g e , di f f er ent but equal or s uper i or ?

    Thi s i s t heref ore the probl em of ack nowl edgement by others t hr ough a re-

    evaluat ion of att ri but es; i t al so appli es to t he previ ousl y dis cuss ed soluti on

    (ii) in whic h t he inferi or group i nvest s i ts al r eady exi st ing separat e charact er

    i st i cs wi t h a new si gnif i cance. A good exampl e can be found in fi eld exper i

    ment s r eport ed by Lemai ne (Lemai ne, 1966; Lemai ne and Kast er szt ei n,

    1971- 2). In one of t he s t udi es , a compet i t i on t o bui l d hut s was ar r anged bet ween

    t wo gr oups of boys at a s ummer camp- but one group was gi ven l es s adequat e

    bui l ding mat eri al s than t he ot her. Both gr oups wer e awar e of t he dis cr epancy

    whi c h wa s bas e d on an ex pl i ci t l y r andom di st r i but i on of r es our ces be t ween

    t hem. The "i nf er i or " gr oup consequentl y engage d in t wo s equences of

    hehavi our : f i r st , t hey buil t an i nfer ior hut but surr ounded it wi t h a s mal l

    gar den; and t hen t hey "enga ge d in sharp di s cus s i ons wi t h t he chil dr en from

    t he other gr oup and the adul t j udges t o obt ain an ack nowl edgement of the

    l egit i macy of t heir wor k. Their ar gument s wer e approxi mat el y as fol l ows:

    we ar e wil l i ng t o admit t hat t he ot her s ha v e buil t a hut and t hat t hei r hut i s

    bett er than our s; but it must equal l y be admi t t ed t hat our smal l gar den wi t h

    it s f ence s ur r oundi ng t he hut i s al so a part of t he hut and t hat we are cl earl y

    superi or on t his crit erion of compar i s on." (Lemaine and Kast er szt ei n, 1971- 2,

    p. 675 , t r ansl ated from t he Fr ench).

    The import ance of the s econd st age, as j ust di s c us s ed, fr om t he point of

    v i ew of our t heoret ical argument is that it s considerat i on l eads t o some cr ucial

    predi ct i ons about int ergroup behavi our. The batt l e for l egi t i macy, in which

    Lemaine's subjects engaged, is a battle for the acceptance by others of new

    forms of inter group compari son. As l ong as t hese are not consensual l y

    ac ce pt ed, t he new charact eri st i cs (or t he re- evaluat ion of t he ol d ones) cannot

    be ful l y adequat e in t heir funct ion of bui l di ng a new social i dent i t y. At t he

    same t i me, t here wi l l be many i nst ances i n whic h the superi or gr oup, for t he

    sak e of its own s ocial i dent i t y, cannot accept one of t he t hree forms of change

    dis cuss ed above: admi ss i on (i) t hat , despi t e t he pr eviousl y exist ing stereo-

    - 26 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    27/64

    - 27 -

    t ype s , t he inferi or group pos s es s es some of the common att ri but es whi ch are

    highl y val ued; (ii) that it s ol d di st i nct iv e att ri but es are at t he posi t i ve end

    of a val ued dimens i on; and (111) that i t s newl y emer gi ng att r ibut es ar e of a

    kind t hat shoul d be posi t i vel y eval uate d. It i s at t his point of t he confl ict

    between comparative social identit ies that the causal processes discussed

    here l ead to t he pr edi ct i on of i nt ense hosti l it y i n intergr oup att i t udes a nd of

    marked discrimination in intergroup behaviour.

    An at t empt wa s made i n t he pr es ent s ect i on of t hi s pr opos al t o pr ov i de a

    l ink bet ween the general t heoret ical formul ati ons of t he i nt roductory sect i on

    and t he r esear ch pl ans of t he sect i on to f ol l ow. Thi s l ink was sought i n a

    consi der ati on, i n part specul at i ve, of the postulat ed proces ses as t hey might

    operate in concr et e soci al contex t s. Before proceedi ng to a di s cuss i on of t he

    r esear ch pl ans , it may be wort hwhi l e to emphas i ze cl earl y the di f fer ences

    bet ween int ergroup pro ces ses ari si ng from an "obje ct i ve" confl ict of goal s

    (as in Sherif's work and related approaches to the problem) and intergroup

    compet i t ion, as dis cuss ed her e, which deri ves from proces ses of social

    compar i s on. Thi s has been succi nct l y done in a recent paper by Tur ner (1973a);

    "We have been led to processes of competition that might be expected

    to operate in many intergroup situations but do not require conditions definedUr-

    by r eali st i c int erest r confl ict - ad ri valr y for a materi al rewar d or goal . (Of

    course positive social identity may be termed a goal but the point is that this

    competition requires no goal which is desirable to both groups of itself and

    woul d ac t a s a r ewar d i n s ome non- i nt er gr oup s i t uat i on). The processes of

    competi t ion whi ch f ol l ow from t he noti on of s ocial ident it y - hereafter cal l ed

    'soci al compet i t ion' - ari se fr om t he ver y nature of an int ergroup sit uati onwher e 'c ompa r abl e' ac t i on i s po s s i bl e t hat i s as s oc i at ed wit h a s har ed v al ue

    dif f eren t ia l .

    To conclude this introductory discussion it is worth making an explicit

    di st incti on bet ween four kinds of int ergroup compet i t i on. This cl assi fi cat i on

    wi l l be hel pf ul f r om t he s t andpoi nt of g aug i ng ho w muc h of an ex pl anat or y

    burden soci al compet i t on may be r equi r ed t o carr y. Fir st l y, t here i s compet i t i on

    whi c h i s char ac t er i ze d pr i mar i l y by t he i ndepe ndent de s i r es of var i ous gr oups

    for a mater ial r ewar d whi ch can be gai ned by onl y one gr oup. 'Mat er i al ' is not

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    28/64

    I

    meant narr owl y; it coul d, for ex ampl e, encompass such t hi ngs as cont rol of a

    poli t ical or social insti t ut ion. The t heme of t his competi t ion i s ex pr ess ed by

    t he noti on of a 'confl ict of i nt er est s'. At t he other pol e, t here i s what we have

    referr ed t o as 'soci al compet i t i on', ari si ng from t he social compar at iv e aspec t s

    of social ident it y as t hey int eract wit h shared va l ues. I ts general l y neces sar y

    condit ions (t heir s uf f i ci ency is a l arger probl em) ar e the sal i ence of t he int er

    group s ituation and the possibi l i ty of differential ly valued actions relevant to

    t he part icular soci al cat egor izat i on (i .e . for our pur poses i nt o 'gr oups').

    The third t ype of competi t ion is defi ned by t he overl ap bet ween the first

    t wo: where a mater ial r eward to some ext ent v al ued of it sel f ser ves as a t oken

    or s ymbol of a val ue dif ferent ial ass oci at ed wit h a possi ble social compari son

    bet ween gr oups. It is an open question whether competition in this case has

    its own distinct behavioural repercussions or whether this kind of situation

    t ends to col l apse int o one of the ot her t wo t y pes , dependi ng on for i nst ance

    j ust ho w muc h i nde pe nde nt v al ue i s po s s e s s e d by t he t ok en or t he de gr ee of

    arbit rar iness i n t he rel ati on bet ween sy mbol a nd sy mbol i zed. It is of interest

    beca use intuit ivel y many worker s t end t o as s ume that t he use of , for ex ampl e,

    smal l monetar y r ewar ds produces a confl ict of int erest sit uati on wher eas infact t heir result s oft en seem more i nt ell igi bl e if one as s umes t hat t he r eward

    had an especi al effect onl y i n so far i s i t hel ped to make sal ient t he possibi l i t y

    of what is properly called social competition.

    A f ourt h t y pe of compet i t i on or compet i t i v e si t uat i on i s wor t h s ug ges t i ng

    al t hough it wi l l not be di sc uss ed in any detail, ' it is t he count erpart of t he

    third t ype in t hat it is defi ned to some ext ent by an overl ap bet ween the first

    t wo and is pr esumabl y a form of t r ansi t i on bet ween t hem. It dif fer s tn t he

    dir ecti on of t r ansi t ion, i .e . in this cas e i t is a social- competi t ive s it uation

    t hat gi ves r i se to a confl ict of i nt erest . Thi s might happen when compar i son

    r esul t s i n a st abl e and expl ici t i nequit y bet ween t wo gr oups and thus t he

    desi r e for posi t iv e sel f- eval uati on l eads to dir ect l y confl ict ing group int erests

    wit h r eg ar d t o t he mai nt enanc e of t he compar at i v e s i t uat i on a s a whol e .

    Ar t i cl es by Thi baut (1950) , Kel l ey (1951) , and Manh e i m (1960) pr ov i de s ome

    support for t he idea t hat such a st abl e inequit y i s an i mport ant factor aff ecti ng

    t he devel opment of s ocial competi t ion into int ergroup host i l i t y." (Tur ner, 1973a,

    - 28 -

  • 8/11/2019 Manuscripts 3 Social Identity, Social Categorization and Social Comparison in Intergroup Behaviour

    29/64

    III . Resea r ch pl ans and met hods

    The test ing of the hypot heses deri ving from t he di sc ussi on pr esent ed in

    t he two preceding sect i ons of t his pr oposal i mposes t wo prel i minary r equir e

    ment s whi ch are int erdependent and need t o be bri efl y di s c us s ed. They are:

    act i ng i n t erms of group rat her t han i n t erms of sel f ; and a cl ear dichot omi zat i on

    by the act i ng i ndivi dual of hi s social worl d into non- overl apping s ocial g r oups.

    The behaviour of each individual can be seen as varying on a continuum,

    one ext r eme cf v/hich consi st s of acti ng full y in t erms of sel f and t he other of

    act ing full y in t erms of his g r oup. These are t heoret ical ext r emes whi ch are

    probably never reached in actual behaviour since it is difficult to conceive, on

    t he one hand, of any social act whic h woul d not be t o some ext ent af f ect ed by

    an indiv i dual 's member ship of vari ous soci al cat egori es and their relat ion to

    t he soci al cat egori es of t hose wit h whom he inter acts ; on t he ot her hand, any

    act i on undert aken in t erms of group members hip wil l a l way s i ncl ude some

    aspect s speci fi c to t he indi vi dual 's own backgr ound, ai ms a nd moti ves whic h

    cannot be f uUy i dent i f i ed wi t h t he inter est s of t he group in t erms of whi ch he

    i s a c t i n g .

    In the experiments described in the previous sections of this proposal(Tajf el et a l . , 1971, and t he subsequent st udies) , act i ng i n t erms of group was

    ac hie ved by the simple devi ce of el iminati ng r es ponse in t erms of self t hrough

    r equest i ng the subject s to divi de bonuses bet ween t wo ot her peopl e. The

    discriminating nature of the subjects ' response, based as they were on tr ivial

    or random criteria of social categorization, can thus be attributed to the nature

    of t he experi ment al si t uat ion. Or ,as Sol e et a l .(1973) r ecent l y wrot e: "Wh y ,

    t hen, on t he basi s of what woul d cert ainl y appear to be t ri vial simil arit y and

    diss i mil ar i t y, do Tajf el 's subj ect s dichotomi ze their worl ds to form "we " and

    "t hey" gr oups? Why do Taj f el 's subj ect s engage in social di scr imi nat ion?

    At f i rst g l anc e it woul d s eem pl aus i bl e enoug h t o a s s ume t hat per cept ual

    ability in inking quantities of dots is a trivial criterion for category membership.

    But consi der once agai n t he subj ect In Tajf el 's experi ment s: he fi nds himsel f

    in a si t uati on whi ch i s , at o nce, social l y impover i shed and const rai ning of his

    behavi our. He is gi ven preci ous lit tl e on whic h he may bas e his dec i s i on, and

    he is f orced t o use what he is g i v en. The crucial i ss ue here i s that to t he

    - 29 -