171

BritishJournal of (2002), 41, 171– 188 © 2002The BritishPsychological Society www.bps.org.uk

HenriTajfel’ s ‘Cognitive aspectsof ’ andthe psychology of bigotry

Michael Billig* Department ofSocial Sciences, LoughboroughUniversity, UK

Thispaper pays tribute to Tajfel’ s classicarticle ‘Cognitiveaspects of prejudice’ and re-examines itscentral arguments. Tajfel’ s paperis important for outlining a social cognitiveapproach to the study of prejudice and also for refuting of whatTajfel called the‘ blood-and-guts’approach. Taking Tajfel’ s propositionthat social psychology is not value-free, thecurrent paper examines the moral andpolitical view of‘ Cognitive aspects’and also the gaps in its approach to the study of prejudice. It is suggested that thiscognitive approach has difficulty in accounting for extreme bigotry,at least withoutrecourse to the motivational themes that the approach seeks toexclude. In particular,there wouldbe limitations in applying this approach in order tounderstand theHolocaust. Indeed, Tajfel did notattempt to do so, for reasons that are discussed. Tajfel’s SocialIdentity Theory (SIT) has similar limitations.The paper also examines Tajfel’s useof the term ‘depersonalization’, whichhe described asa ‘milder’form of dehumanizationof out-groups. Later socialidentity theorists have tended touse ‘depersonalization’differently, shifting their attention to in-groups. Their perspective movesaway from understandingthe topic of prejudice in the way thatcan be foundin Tajfel’s ‘Cognitiveaspects of prejudice’ . Finally,the present paper suggests how extreme prejudicemight be studied without returning to the motivational ‘ blood-and- guts’approach that Tajfel so cogently criticized.

Itwillsoon be 20yearssince the death of HenriTajfel. H isinflue ncethroughout social psychologypersists, e speciallyin work on socialidentity (Robinson, 1996).A sBrown and Capozza(2000) have shown, interest inSocial I dentity Theory (SIT)continuesto grow,with anincreasingnumber of studies beingpublished yearly.Aswith any major figurein the socialscie nces,T ajfel’s writingsrepay careful study and reinterpretation. Thisarticle ex aminesthe rhetoricand argument of HenriTajfel’ s classicarticle ‘Cognitiveaspec ts of prejudice’,whichwas first published inthe Journalof Biosocial Sciences in1 969and which has beenreprintedanumberof timessince.Thearticle featuredasakeychapter inT ajfel’s book Human groupsand social categories (Tajfel, 1981). Thereare a numberof reasons for re-examining‘ Cognitiveaspec ts of prejudice’. Most notably,the articlere mainsone of the most brilliantpiec esof writingin the history of socialpsy chology.Writtenbefore socialpsy chology’scognitiveturn inthe

*Requestsfor reprints should be addressed to Michael Billig, Department of SocialSciences, Loughborough University, LoughboroughLE11 3TU, UK(e-mail:[email protected]). 172 MichaelBillig

1980s,‘ Cognitiveaspec ts of prejudice’outlined the principlesof cognitivesocial psychologywithout specificallyusing the term ‘socialc ognition’. Eiser(1996)has arguedthat current assessments of Tajfel’s legacytend to concentrate on SocialI den- tity Theory at the expense of hise arlierwork insoc ialjudgment. T ajfelproduc ed ‘Cognitiveaspec ts’at akeypoint inhis inte llectualdevelopme nt. Itwas written asa culminationof hiswork insocial judgme nt, but before he developed SocialI dentity Theory.The articlewas publishe djust asT ajfelwasembarking on the minimal intergroup experimentsthat wereto laythe basisfor hislater intergroup theorizing (Tajfel,1970; T ajfelBillig,B undy,&Flament,1971). A ssuch,‘ Cognitiveaspe cts’ combinesthe themesof socialjudgment with those of intergroup conflict, asTajfel arguedthat the principlesof cognitioncan illuminate the psychologicalnature of prejudice. Overand abovepay ingtribute to Tajfel’s enduringinte llectuallegacy ,there is another reason for returningto ‘Cognitiveaspe cts of prejudice’.Thisis to understand the nature of prejudice.Itisnot suggestedthat ‘Cognitiveaspe cts’holds allthe keys, nor eventhat weshould follow rigidlythe messagethat Tajfelwas advocatingthere . Farfrom it,the strategy isto examinec riticallythe omissionsin the article.Thisis not to downgrade the importanceof ‘Cognitiveaspe cts’. In the socialsc iences,works of the highestintellec tualquality alway srepay criticalre- examination,for eventhe ir limitationsand omissionscan berevealing.T hisis what separatesthe first-ratefrom the merelycompe tent. The omissionsin ‘ Cognitiveaspe cts’, itwill be suggested,enable us to understand the contextof Tajfel’s theorizing,espec iallyin relation to hisbac kground. They are alsothe oreticallyre vealing,inasmuch as they point towards the limitationsof the cognitiveapproac hfor understanding extremeprejudic e.Inthis respect, the present tribute to ‘Cognitiveaspe cts’is also an argument.To argue, howe ver,is not necessarily to reject, but to develop.A sTajfel(1981) e mphasized,no socialpsy chologyis value - free,for allsocial psy chologyreflec ts the culturalc limatein which itis produced. That was why he arguedthat itwas vitalto havemultiple pe rspectivesinsocial psy chology (Tajfel,1981, p. 6).The political climate of today isdiffe rent from that of the late196 0s and early1 970s,when Tajfelwas developinghis ide as.So must socialpsy chology reflectthesechanged times. T oday’ssocialpsy chologyc annot be amerere petition of that whichwas formulated ageneration ago.Tajfel may have argue dthat motivational themesshould be put to one side,but, by consideringone of the finest piecesof writing,it will be suggestedthat suchthemescannot, and should not, be excluded entirelyif one wishesto understand extremepre judice.

Imageof humanity Theopening paragraphs of ‘Cognitiveaspe cts’illustrate Tajfel’ s intellectualsty le.He does not start with acareful reviewof previousex perimentalstudie son prejudiceand socialjudgment. That sort of reviewcan be found inhis contribution to Thehandbook ofsocial psychology ,whichwas publishedinthe samey ear(Tajfel,1969b) but which Tajfeldid not includein Human groupsand social categories .Tajfelnot only read widelyin othe rdisciplinessuch ashistory ,anthropology and politicalsc ience,but also usedthis readingin his social psy chology(se e,for instance,hiscomme nt inthe interviewwith Cohen(1977,p. 306)about hisreading). The M arxist philosopherand cricketwriterC.L.R.Jamesadaptedasayingfrom Rudyard Kiplingin the prefaceto Cognitiveaspects of prejudice 173 his book Beyond aboundary :‘‘What do theyknow of cricketwho only cricket know?’’(James,1964, p. 11).A further adaptation would fit Tajfel’s own academic practice,aswe llas hisinte llectualposition: ‘‘What do they know of socialpsy chology who only socialpsy chologyknow? ’’ ‘Cognitiveaspe cts’begins with socialanthropology and an assertion of the com- monalityof humans. Tajfelwrites that the viewsof earlysocial anthropologists suchas Rivers,who believedinthe culturaland psychologicaldefe ctivenessof non-Europeans, arenow rejected. Tajfelillustrated this with alongquotation from Leévi-Strauss’s Thesavage mind .Inthe quotation, Leévi-Strausswas arguing that so-calledprimitive cultures havedeve loped asophisticated practicalscie ntificknowledge. A llforms of agriculture,the makingof pots, the use of poisons inwarfare or ritual,and so on, depend upon an empiricalattitude and adesirefor knowledge.Tajfelthen commented that: ‘‘animageof Manemergesfrom theseconsiderations’ ’ (1981,p. 128;quotations aretake nfrom the version of ‘Cognitiveaspec ts’included in Human groups ; here, and insubse quent quotations, no attempt willbe madeto alterTajfel’s use of masculine nouns and pronouns to denote humanity ingene ral).The image depic ts humans as beingessentially rational be ings,who seek,above all, to understand their world.The sameimage, T ajfelcomments,is to be found inthe work of FrederickB artlett, who stressedthe human ‘effort aftermeaning’. AsTajfelre alized,such imagesof humans cannot be provedby experimentation. Evidencec anbe brought from experimentalpsy chology,and from history,socialanthro- pologyand so on to support the image,but ultimately,the imageste msfrom awider politicaland moralperspec tive.B ecausesocial psy chologyand othersocialscie ncesare not value-free,such animagewill conve ymoraland politicalvalue s.I nthis way,Tajfelat the start of ‘Cognitiveaspec ts’was dec laringa politicaland moralposition, just as Leévi-Strausswas doingwhe nhe appliedthe rationalimage ,not to modern Europeans, but to the verycultures that Europeans had for centuries dismissedasirrational. Typically,an academicargument is dire cted towards arivalposition. Therefore, to understand the meaningof an academicposition, not only must one determinethe specificcase that isbe ingadvocate d,but one must alsoknow what positions arebeing arguedagainst (B illig,1 987a).T ajfel’s assertion of acommon rationalhumanity was an argumentagainst a counter-position. Ashe noted, when observers try to explain human socialactivity ,especiallywarfare, they discard the rationalimage of humanity, adopting instead ‘‘ablood-and-guts modelfor socialphe nomena’’ (Tajfel,1981, p.128).D uringthe 1960s,the blood-and-guts modelof conflictwasprominent inboth popular and academicwriting: it had been‘‘blessedand speeded on its way by a numberof books, someof whichhave quic klybec omebest- sellers’’ (p. 129).T ajfel had inmind populist versionsof Freudianand ethologicaltheories that postulatedan instinctfor aggression.F oremost amongstthe blood-and-guts writers was the Nobel prize-winningethologist, K onrad Lorenz,who was to publish hiside asin a numberof best-sellingbooks (e.g.L orenz,197 4,1976). L orenzdirec tly extrapolated from his animalstudie sto human beings,in way sthat anticipatesome of the cruder versionsof today’sevolutionarypsy chology.Lorenzc laimedthat humans, incommon with a numberof animalspe cies,had evolvedan aggressiveinstinc tbut without the biologi- callybase dinhibitory mechanismsto be found inothe rspecies.L orenzpropose dthat human conflict should be explainedin terms of this innateaggre ssiveinstinct, and, thus, hewas proposing abiologicalsourc eof irrationalityat the coreof human nature. Inthis regard,L orenz’s blood-and-guts modelwas the veryantithesis of the rational modelto be found inB artlett and Leévi-Strauss. 174 MichaelBillig

Againstthe blood-and-guts model Tajfelrec ognizedthat therewe recompe llingintellec tualand politicalre asons for combatingthe blood-and-guts model.In ‘Cognitiveaspe cts’, Tajfelproduc edtwo sets of argumentagainst the blood-and-guts model.F irst,the rewas the waxing-and-waning argument, whichpowerfully yet simplyunde rminedthe theoreticaladequacy of an instinctiviste xplanationof human warfare.T ajfelconcededthat ‘‘allme ncanand do displayhostility towards groups other than their own’’ (1981,p. 130).H owever,the y do not do this allthe time,for ‘‘there isalso no doubt, however,that under other conditions this hostility eitherdoes not appearor can be modified’’ (p. 130).I nshort, the positingof aninvariantand unchangeableinstinc tcannot explainthe waxingand waningof socialc onflict(nor of individualaggre ssiveness). Toexplainwhy wars occur at one timeand not another,one must gobe yond the hypothesisof an aggressive instinctand look at socialand historicalconditions. Thisinvolves e xaminingthe beliefs and ideologiesthat groups hold about eachother. Thus, the study of conflictmust embracethe study of group attitudes, and,in conse quence,the reshould be asocial psychologicaldimension. Tajfelalsohad amoral/politicalargumentagainst the blood-and-guts perspective.T o offeran account of socialc onflictinterms of an unchanginginstinc tis,at best, to suggest that nothing canbe done to alleviateprejudic e.A tworst, itis to justify prejudiceand chauvinismas an innatepart of the human condition. AsTajfelwell knew,this was acentralpart of Naziideology ,whichpostulatedthat thereis an innate desireto favour one’sown race.A ccordingto this ideology,liberalismand abeliefin the universalityof human rationalitywe re‘ alien’to the nature of humans. Atthe time of writing‘ Cognitiveaspec ts’, Tajfelwas unawarethat Lorenz’s earlypape rs,which werepublished duringthe Naziera, were lac ed with Naziterminology ,evento the extent of praisingthe racialpolic yof the Naziregime (see ,for instance,Lorenz,1940, 1943).T he exposure of Lorenz’s earlywork wasto comeshortly after the publication of ‘Cognitiveaspec ts’(see Eisenberg,1972; se ealsoBillig, 1978; N isbett, 1976). Evenwithout the specifice videncelinking L orenz’s instictivismwith Naziide ology, Tajfelwas aware of the deeplyreactionary implic ations of the blood-and-guts approach. Inthis light,his use of Leévi-Strauss to beginhis analy sisof prejudicewas no accident.I twasbetter to beginwith the assumption of auniversal,human rationality, asL eévi-Strauss was doing—and, indeed,as C homsky was doingat that time—than to utter profound-sounding banalitiesabout the irrationalityof human nature.

Cognitionand prejudice Thene xtstep inT ajfel’s argumentwas a bold one:to show that the seemingirrational- itiesin human socialc onduct owed theirorigin to this essentialrationality .Again,the argument issimple and allthe morepowerful for itssimplic ity.Tajfelstarte dfrom the assumption that prejudiceconstituted aserious socialproble m.I ndeed,as he recounted in Human groupand social categories ,the essaythat won hima scholar- ship to enter universityas a mature student afterthe warwas entitled‘ Prejudice’ (Tajfel,1981, p. 2).I n‘Cognitiveaspec ts’, Tajfelwrote that prejudicelite rallyme ans ‘prejudgment’, and most commonly,the termis used to referto the prejudgment of other groups.Technic ally,itis possible to prejudgein favour of othergroups.H ow- ever,Tajfe l’s concern,whic hisre flected inthe common use of ‘prejudice’to denote a serioussocial problem, was the prejudiceagainstother groups.A she wrote, the Cognitiveaspects of prejudice 175 purpose of ‘Cognitiveaspe cts’was ‘‘to present an outlineof the cognitivee tiologyof prejudice,mainlywith regardto its unfavourable aspects’’ (Tajfel,1981, p. 131, emphasisin original). Inprejudicedthinking,judgme nts aremade about the membersof other groups regardless of theirindividual charac teristics:members of the out-group arejudged negatively,or unfavourably stereotyped,simply bec ausethey belongto the out-group. Tajfelrelatedthis type of stereotypingto ordinary sense-making.I norder to under- stand the world—both the physicaland the socialworld— humans needto make cognitiveshort cuts. Thereis too much sensory information availableat any one point to dealwith every detail.U nlessthis information isc ognitivelyorganized— unle ss itis categorized—there can be no meaning.T herefore,humans needto organizethe social world into categories.H owever,whe nwedo this,we are liable to distort the world, evenas we make it meaningful. A sTajfelhad shown inhis line -estimationstudies, imposinga categorizationon acontinuum of stimulicreate satendencytowards two sorts of exaggeration:the reis the tendencyto overestimatethe extent to which instancesof the samecate goryresemble each other and atendencyto overestimatethe differencesbetween instancesof different categories(Tajfel&Wilkes,1963). Tajfe l arguedthat the effects of categorizationon the judgmentof physicalstimuli re sembled the exaggerations of socialstereoty pes.T hus, the tendency to prejudgemembers of out-groups was,at root, similarto the moregene ralte ndency to exaggeratethe differencesbetweencategoriesand to minimizediffe renceswithin categories. The notion of ‘categorization’was one of three keyconcepts that Tajfelused in ‘Cognitiveaspec ts’to outlinethe cognitivedy namicsof prejudice.T he other two concepts were‘ assimilation’and ‘coherence’.Inlinewith Tajfel’s attempt to construct agenuinelysocial approach to prejudice,he stressedthe importanceof ‘assimilation’. Individualsdo not createtheir own categoriesbut assimilatethe categoriesthat areculturally available ,thereby acceptingc ulturallyde terminedpatterns of prejudgment and stereotyping. Individualsuse thesesocial cate goriesto makesense of, and thereby bring coherenceto, theirunderstanding of the world.T he searchfor coherenceprovidesa clueabout how individualsc ope with understandingthe constantly changingsocial world.I ndividualswill atte mpt to use categoriesinway sthat preservethe irse lf-image or integrity.Accordingto Tajfel,‘ ‘this needto preservethe integrityor the self-image isthe only motivationalassumption that wenee dto makein order to understand the direction that the searchfor coherencewill take’ ’ (1981,p. 137).A similar motivationalassumption wasto appearin Social I dentity Theory,whichassume da needfor apositivesocial ide ntity (i.e.Tajfel&Turner, 1979). ‘Cognitiveaspec ts’provide sabrilliantre buttal of an instinctualtheory of prejudice. Tajfeltakesan aspect of human behaviourthat appearsto be inherently irrational but arguesthat this irrationalityshould be understood interms of apsychological perspectivethat isbase don the assumption of human rationality.Thereis, in c onse- quence,no need to posit an underlyingmotivational force .Yet,as c an be seen, motivationalpre misesarenot entirelye xcluded.The cognitive approach includesthe motivationto understand, aswellas the motivefor preservingthe self-image.A ssuch, the cognitiveapproac hisnot entirelycognitive. Thereare a number of criticismsthat can be madeof the cognitiveapproac hto prejudice.Recently,criticshave suggeste dthat cognitivesoc ialpsy chologyis mistaken intaking a perceptual,rather than discursive,view of categorization(B illig,1985, 1987a;Potter &Wetherell,1987; Wethe rell,1996). A ccordingto this argument, the 176 MichaelBillig categoriesof prejudicearee ssentiallylanguage c ategories.Spe akers canuse language flexiblyand, thus, arenot restrictedmerelyto minimizingwithin- categorydiffe rences or between-categorysimilaritie s(Edwards, 1991).I ndeed,with language,we can both categorizein our judgmentsas we llas particularize,not to mention talkc riticallyabout our categories(Billig,1987a). A llthis makesthe use of categoriesinlanguage ve ry different from the use of perceptual categorization.M oreover,this dualityof particular- izationand categorizationc an be found inthe languageof those who areprejudic ed and,more ge nerally,inthe talkof those who hold strong views(Billig,1985, 1991). The present purpose isnot to pursue this discursivecritique of the cognitive approach. Instead,the tacticis to apply Tajfel’s theoreticalcritique against the blood- and-guts approach to hisown approach. Inthe waxing-and-waningargume nt, Tajfel was exposingwhat the blood-and-guts approachcannot explain.The fact that the argument could expose the limitationsof the blood-and-guts approach does not mean that the cognitiveapproach has no analogouslimitations. I nfact, itis safe to assume that any theory insocial psy chologymust be incomplete,just asany imageof the person that underliesatheoreticalpe rspectivemust alsobe anoversimplification.The intention isto probe the absences,in orde rto clarifywhich aspects of prejudice the cognitiveapproach best addressesand whichaspects ittends to ignore.

Holocaustand explanation Thereis a paradoxin the limitationof Tajfel’s approach to prejudice.Hedid not apply either hisc ognitiveapproach, or hisSocial I dentity Theory,to explainthe one histori- cale ventthat brought himto socialpsy chology—.I nthe openingpages of Human groupsand social categories ,Tajfeldescribesin all too briefterms his passagetowards socialpsy chology.Unlikethe majorityof the peoplewhom he had grown up with, Tajfelsurvive dthe massacreof European Jews.F or the rest of hislife , he wasto reflecton what had happened. Ashas been mentioned, Tajfel’s essay ‘Prejudice’enabledhim to enter university asamature student, and he was to become aprofessional academic‘ ‘almostin a fit of absent-mindedness’’ (Tajfel,19 81,p. 1;see alsohis ac count inthe interviewinC ohen,1 977).A lthough Tajfelmay have be gunhis researchcaree rby addressingtec hnicalque stions of perceptual judgment,he was soon usingthis researchasa basisfor understanding the nature of prejudice.Academic researchfor its own sakewas insufficie nt for Tajfel.H ewould tellhis students that behind hiswork layone question: how isge nocidepossible? Yet,apart from occasionalcomme nts, Tajfelrarely in his writte nwork appliedhis powerful theoriesof prejudiceto the one eventthat preoccupied him(for amore detaileddiscussion of this omission,se eBillig,1996). I npart, this reflects amore generalphenome non. Therewas littlewritte nor spokenabout the Holocaust for a generation afterthe Second World War.Survivors at that timewould rarelyrisk tellingthe irstorie sto aworld that seeminglyhad moreimportant thingsto do than listen.I naddition there arespec ificreasons againstoffering academice xplanations, especiallypsy chologicalex planations,of the Holocaust. Ifwec laimto haveex plainedan event,we are claiming to haveunder- stood it.M oreover,the explanationby providingunderstanding— by givinga set of adequate causes—see msto ‘wrap up’that event.F urtherdetailedthought becomes unnecessary;afterall,we know and understand what happened.A ny explanation of the Holocaust would be claimingtoo much.M ore than any otherevent inhistory ,the Holocaust should not be explainedaway;itdemands furtherthought. Apsychological Cognitiveaspects of prejudice 177 explanationruns afurtherrisk: toutcomprendre c’est toutpardonner .Ifan account explainsthe motivesand thought processesof the perpetrators inpsy chologicalte rms, suchanaccount maypre sent the perpetrators inan understandable lightthat wittingly or unwittinglyinvites e mpathy (see Billig,1996; M andel,1998). P rimoL evie xpressed this better than anyone in If this is aman (Levi,1987). H earguedthat one should not understand what happened inthe Holocaust becauseto understand isalmost to justify. Understandingwould meanputting oneselfin the placeof, and thus identifyingwith, the perpetrators. Better not to understand, arguedLevi,than risksuch identification. Inaddition to such generalc onsiderations, there arespec ificfeature sof Tajfel’s cognitiveapproac hthat would makeit unsuitable, at leaston its own, asan expla- nation of the Holocaust. In‘Cognitiveaspe cts’, Tajfelwas describingunive rsal processes.‘ Categorization’, ‘assimilation’and ‘coherence’werenot assumedto be culturallyspe cific,as if the yonly occur inc ertainsocio- historicc ontexts and not others. Theywereinte nded to represent featuresof allhuman thinking.Similarly , Tajfelwas proposing that itwas human nature to wish to understand the human world and that the simplificationsof categorizationwere ne cessary to do so. Theseconce pts can be appliedto allforms of intergroup prejudice.However,if the yareapplied to the Holocaust asifoffering anexplanation,the extremespec ificityof that eventwould be lost. Itwould be neitherappropriate, nor informative,to say that the Germanssy stemati- callymurdere dthe Jews inthe Second World War,bec ausethey weresee kingto understand the world and to protecttheir self-integrity.Ofcourse,such processesmay haveplay edtheirpart, but constitute only asmallpart ina widerpicture (Mandel, 1998,200 1).W ecan say—indeedwemust say—that the Germanmurder of the Jews should be understood interms of what the Nazisthought about Jews.B ut this isvery different from offeringan ‘explanation’in te rmsof cognitiveproce ssessuchasthe searchfor understanding and the protection of self-integrity.Thereis nothing intrinsic about these processes,nor about the universalproce sses of categorization,assimilation and coherence,that would account for the historicalspe cificityof Naziide ology. Itissignificant that Tajfelpresented hiscognitive approach asoffering an under- standing(not an explanation) of prejudice.One implication of the approachisthat prejudice,at leastin its literalse nse of prejudgment,is ine vitable:we must use cat- egoriesinour thinking,and categorizationpredisposes us to prejudgment. Thus the inevitabilityof prejudice,and indeedof stereotyping,is implied (see Billig,1985, 1987a,for criticismsof this assumption). Evenif it is c oncededthat prejudiceisine vitableand that human thinkingabout socialgroups involvessome or other form of stereotyping,this does not mean that all prejudicesand allstere otypingare equivalent. I ndeed,the term‘ prejudice’may be too anodyne to coverall forms of intergroup stereotyping.Ste reotypes,e venifthey are broadly ‘negative’,can be distinguishedin te rmsof theirinte nsity and ideological importance.For instance,the current Britishstereoty pe of Germantourists, who rush to the beachto secure the best sunbathing spots, can be distinguishedfrom the Nazi of Jews asvermin,responsible for allG ermanmisfortunes. AsTajfelnoted in‘ Cognitiveaspec ts’, particularcate gorizationsare embedde d(or assimilated)into widerideologicalpatte rns of belief. The term‘ prejudice’isnot aneutral term,but itconvey sacritiqueof the beliefs labelledas ‘ prejudiced’(see Billig,1991). D espite this moraland politicalevaluation, not allprejudic es,just likeall ste reotypes,nee dto be considered asequivalent.Within the generalcate goryof illegitimatebelie fs, ‘prejudice’isnot the strongest termin the 178 MichaelBillig lexicon.The reis, for instance,‘ bigotry’.In extremeideological conte xts, ‘prejudice’ can be consideredacomparativelymild te rmand, indee d,inappropriate becauseof this comparativemildne ss.To say that Hitlerand other leadingN aziswe re‘ prejudiced’ againstJe ws would seemto be an understatement. Theiranti- semitismcrie sout for a strongerterm.T hus, one needs to intensity the lexiconin orde rto beginto capture the sort of beliefs that layat the coreof the sort of Naziide ology.Aterm that describesa phenomenon, that issuppose dly universalto allbelief sy stems,will not do for abelief systemthat isspec ificallyge nocidal,or ‘eliminationist’, to quote Goldhagen(1998). Tajfel,with hisfine sense of understanding and appropriateness,would not have neededto be told this.I nthis lightit is not surprisingthat he did not turn his theoreticalc onstructs to the very issuethat ledhim to the study of ‘prejudice’.

Prejudiceand bigotry Tajfel’s argumentagainst the blood-and-guts approachcanbe turned around and appliedagainst his own approach. Just asthe blood-and-guts position could not account for the waxingand waningof warfare,so wec an say that therec an be a waxingand waningof prejudice.Sometimesthere issoc iallyshared bigotry;sometimes thereis not; sometimesan ideologyof tolerancemightbe widespread.A cognitive approach that linksprejudic eto categorization(and alsoto assimilationand coher- ence)cannot of itselfacc ount for this waxingand waning:some additional e lement is called for. This,of course, doesnot necessarilyimply a return to the blood-and-guts modelthat Tajfelcriticized so powerfully.Just to say that the cognitivemodel ne eds an added element isnot to say that the additionalfac tor must be anassumedinnateinstinc tfor aggression.I tmightbe suggestedthat the additionalele ments arenot psychological factors but arehistoric aland culturale lements, which,acc ordingto Tajfel’s account, would need to be assimilatedby the individual.I fthis werethe case,the distinguishing features betweendifferent intensitiesof prejudicewould not be psychological.It would be implausible,however,to assert that the differencebe tweenprejudice,asa cognitiveinte rpretation of the socialworld, and bigotry,asan intense group hatred, must only be culturaland historical.T heresurely could be somesoc io-psychological distinctions betweenthe two typesof phenomena. Thedifferencebe tweenprejudiceand bigotry,or hatred, canbe consideredfurther. Whereasthe notion of prejudiceseemsto invitea cognitiveinte rpretation, sinceits literalsense re fers to prejudgment, bigotry seemsto includethe verypsy chological components that the cognitiveapproach sought to exclude,or at leastto put to one side.Theseare emotional or motivationalfac tors, addingan intensity and wilfulness to merecategoric ale xaggerations.T otalkof Nazianti- semitismwithout suchfactors—to presume itto bemerelya categorymistake— would seemodd. Kenneth Burke (1963),inhis book Agrammarof motives ,discussesthe limitations of givingsimple (‘ nothing but’) explanations of human behaviour.T he excluded elements havea habit of wormingtheir way back.B urke specificallymentioned instinctivistex planations of warfare.Whenappliedto actualc ases,suc hexplanations start recognizingsituational fac tors, inorder to account for why the warocc urred at one point and not another(Burke, 1962,p. 34).Similarly ,the cognitiveex pla- nation cannot account for why prejudicemightbecome bigotry without discussing motivationalissue s.O ne mightpredict that the moreit avoids doing so, the morethese issueswill sne ak into the account by,asitwere, the backdoor. Cognitiveaspects of prejudice 179

Thiscan be seenin‘ Cognitiveaspec ts of prejudice’.Atone point, Tajfelwrites ‘‘if amanis pre judiced,he has an emotionalinvestme nt inpre servingdiffere ntiations between hisown group and ‘others’’ ’(p. 134).H ereT ajfelisusing the description ‘prejudiced’in a different waythan he used ‘prejudice’when describingthe cognitive implications of categorization.T obe ‘prejudiced’in this statement isto do morethan use asocialc ategoryin a waythat overestimatesthe differencesbetweenin-group and out-group members.T he sentencesignificantly suggests that prejudicec an be acon- dition of aperson’s being(‘ ‘to be prejudiced’’, he writes), not somethingthat isa by-product of inevitablecognitive processing. I nthis case,the condition of being includesanextra dimension, namelyan ‘emotionalinvestment’ . Tajfelhas littleto say about what this ‘emotionalinvestme nt’might be and how it mightoperate. Therewas a good theoreticalreason why Tajfelwould haveavoided drawingout this theme.Totalk of emotionalmotivations might seem to be leading back to individualdy namics.Significantly ,inthe statement just quoted, Tajfelrefe rs to an individualbe ingpre judiced,and not agroup.R educingsoc iale vents to individual motiveswas somethinghe consistently sought to avoid.T he problemof bigotry was not to be resolvedby the psychologicalanaly sis—or psychoanalysis—of individual bigots. However,the reductionism,that Tajfelopposed, isnot anecessary consequenceof consideringmotivational factors. Recently,within socialconstructionist psychology, there isan awareness that emotions must be considered associally constitute d (Edwards, 1997;H arreé&Gillett,1994). This position stresses that emotions do not exist aswordless impulses,ly ingbe neath sociallife ,but areconstitute dwithin social, discursiveinterac tion. Thisis e ventrue of unconscious emotions (Billig,1999). T hus, hatred neednot be seen asan individualcondition, located within the body of the individual.There can be ideologiesof hatredthat produce‘hate-talk’(e.g. Whilloc k& Slayden,19 95).T hehatred isnot separate from the discourse.Tohate isnot merely,or principally,to feelsomething at abodily or viscerallevel— but to believeand to utter particularsorts of thingsabout others (Billig,2001 ).Without that, there can be no hatred. Toput itcrude ly,wedo not need to know the hidden,inne rpsychological mechanismsof the Nazis,in orde rto know that theyhatedJews.The iractions and their words werenot a sign of their hatred, asifthe hatred reallywas e lsewhere;those actions and words were,in the most literalse nse,pure hatred.

Gap inSocial Identity Theory Although ‘Cognitiveaspec ts’containe dfeaturesthat wereto be developedinthe later SocialI dentity Theory,thereare also differe nces.M ost notably,socialide ntity theory is not atheory of prejudice.It certainlyis not atheory of murderous bigotry.Itis,at root, atheory of group freedom. It tellsof the way that oppressedgroups canfind waysto challengegroups that havethe power to ascribeidentities and .The most originalparts of the theory describehow groups can re-createstereoty pesthat are appliedto them: they canfind newdimensions of comparison,alte rthe valuationof existingtraits, collec tivelyoppose powerful out-groups, etc.Significantly,the exam- plesthat Tajfeltended to use,in order to illustratethe seproce sses,were the black powermovementin the USAand the women’s movement. Both weredire ctlyover- turning ascribedstere otypes and re-creatingthe irsocial ide ntities. Tajfelbelonged to the generation of East European Jewsfor whom attachment to the Jewishstate ,especiallyafter the war,was de eplyhe ld.A fterwhat that generation had 180 MichaelBillig witnessed,a world without aJewish state could neverbe fullytrusted. Nevertheless, with the exception of afew passingrefe rences(e.g.T ajfel,19 81,p.186),Tajfel did not use Zionismas an illustrationof SocialI dentity Theory inthe way that he usedblack power and the women’s movement. Zionismhad been engagedinthe veryac tivity describedby SocialI dentity Theory;itwas explicitlyove rturning centuriesof stereo- typing,which had depicted Jewsaspassive,unheroic ,politicallyuncre ativeand so on. Inprivateconversation, especially with Jews,Tajfel would be open inhis support for the existenceof Israel(whiledistancing himself from the particularactions of Israeli governments). However,in the 1960sand 1970s,the rewe restill inhibitions, at least in Britishunive rsity circles,against ex pressingJewish issue stoo openly.Inwaysthat SocialIde ntity Theory itselfcould describe,there werepressures not to be ‘too Jewish’ too visibly(se eBillig(2000) for an account of how such pressures affected the cultural sphereof popular musicat that time).In the late1960s, espec iallyduring the student power events,Z ionismwas not politicallyac ceptablewithin leftist,liberal circles of the socialscie nces(significantly,Tajfeldisc ussedanti-semitismof both the left and the rightin his interview with DavidC ohen to amuchgreater extent than inhis writings of the time—seeC ohen, 1977,p. 310).O nthe far left, anti-Zionismwas c ommonly expressed(Billig,1987b). The consequenceis that Tajfel’s writingsshow agapat the core of hispolitical thinking.H isc ognitiveapproach did not directly address the extremeof bigotry that constituted .N or did he apply SocialI dentity Theory to Jewish(or Palestinian) identity inthe MiddleE ast.Soc ialI dentity Theory,however,c ould not havebe eneasily usedasthe centrale lement ina socialpsy chologicalinte rpretation of Nazism.A social identity theorist, who wished to apply the theory,mightsay that Nazismarose bec ause the Germanscreated dime nsions of comparisons with the Jews,in orde rto producea positiveG ermanself- identity.Theremight be agrainof truth insuch an account. Indeed,it would not bedifficult to point towards such elements inthe development of Naziideology and eveninthe thinkingof Hitler(Mandel,2001). B ut to leavethe matter therewould convert the grainof truth into awideruntruth. Thisuntruth would be a grossoversimplific ation,whic hfailedto probe the essentialc haracteristicsof Nazism. After all,N azismwas not merelya searchfor positiveG ermanse lf-identity.Much more was at stake. Therewould alsobe apoliticaluntruth inapply inga theory of liberationto the most reactionarypolitical move ment imaginable.Itwould suggest equivalences,which should not be made,suchasapsychologicale quivalencebe tweenthe development of in-group identity inN azismand that of the women’s movementor the blackstruggle for equality.Any empiricallygrounded similaritieswould be trivialas compare dwith the differences.The more that the socialpsy chologicaltheory drew attention to such similarities,the moreinappropriate would be its application.T hisvery inappropriate - ness shows the truth of Tajfel’s assertion that socialpsy chologicalthe oriesare not, and should not be,value- free.

Depersonalizationand dehumanization InSocialI dentity Theory,there isa furtherabsence,parallelingthe cognitive approach’s failureto distinguishbe tweenprejudiceand bigotry.Inhiswritings about socialidentity ,Tajfelintroduce dthe notion of ‘depersonalization’— aconcept that was to feature significantly inT urner’s Self-CategorizationT heory (see,for instance,Turner, Cognitiveaspects of prejudice 181

Hogg,O akes,R eicher,& Wetherell,1987). A parallelcan be drawn betweenTajfel’s comments about depersonalizationand histreatme nt of categorizationin ‘ Cognitive aspects’. Tajfelsugge stedthat the use of categorieswasne cessary for human thinking,but that this entailedprejudgment and,thus, prejudice.H owever,an extra emotional investment was required to turn such inevitablecognitive pre judiceinto the state of ‘beingprejudice d’. Depersonalizationwas similar to categorizationin being a common aspectof intergroup phenomena.Tajfel wrote that depersonalizationocc urredinthe deliberatelyartificial situation of the minimalintergroup experimentsas well as in actualsituations of warfare:‘‘Thecommon denomination of allthese examples—from the contrivedabsurdities of the experiments to the tragicrealities of the War—isthe depersonalizationof the membersof the outgroup’’ (1981,p. 241). Significantly,Tajfel,in his c omments about depersonalization,also use dastronger term,namely ‘ dehumanization’. Depersonalizationmight be the ‘common denomi- nator’in the minimalintergroup experiments and inactual warfare, but thereis a crucialdiffere ncebetween the two situations. Inactualwarfare ,out-group members areofte nnot merelydepersonalize dbut commonly dehumanized.Thus, T ajfelwrote that depersonalizationof out-group members maybe just abeginning,and ‘‘the next stageis often theirdehumanization’ ’ (1981,p. 241).D epersonalization,thus, isa milderform of the way that in-group members can treat out-groups:‘ ‘Our socialhistory isfull of familiarand horrifyingex amplesof dehumanizationof outgroups and even moreso of milderforms of their depersonalization’’ (1981,pp. 52–53).A ssuch,there isa ‘continuum’stre tchingbetwe en depersonalizationand dehumanization(Tajfe l, 1981,p. 241).O ne mightsuggest a ratio:depersonalization is to dehumanizationas cognitiveprejudic eisto bigotry. Just asthe cognitiveapproac hdoesnot stipulatethe factors that leadfrom prejudice to bigotry (or from prejudicesas‘cognitiveprejudgme nt’to prejudiceasan ideologyof hatred), soTajfelin his soc ialide ntity theory did not elaborateon the possiblecontin- uum betweendepersonalizationand dehumanization.F or any socialpsy chological theory of socialc onflict, letalone a theory of genocide,suc hacontinuum would be vital:it would demarcatethe ‘ordinary’from the ‘abnormal’, or the mildfrom the strong. AsTajfelstressed,ideologic alfactors, or socialmy ths, willbe crucialon this continuum, for cruelty and massacreare made le gitimateif the victimsare see nas beingin some way ‘inhuman’(T ajfel,1984, p. 698).Merelyto be categorizedas amemberof an out-group isnot sufficientto be consideredasbe ingnon- human, although,conversely ,itc an be saidthat an ideologyof dehumanizationmust rest on the distinctions of socialc ategorization.

Later work on socialidentity Itmightbe thought that laterwork on SocialI dentity Theory would havediscussed the continuum betweendepersonalizationand dehumanization.T hisis e speciallyso since muchof this laterwork has involvedgiving the concept of depersonalizationa morepromine nt roleand has addressed itselfto makingthe originaltheory more precise.H owever,the rehas beenacrucialc hangein the use of depersonalizationin muchof this laterwork. Thisc hangenot only leadsaway from the problemof distinguishingthe mildfrom the extreme,but alsoleads away from someof the intergroup considerations that underlayTajfel’ s theorizing. 182 MichaelBillig

The concept of depersonalizationis much morepromine nt inT urner’s Self- CategorizationT heory than itis in e itherTajfel’s ‘Cognitiveaspec ts of prejudice’or his socialide ntity theory.Turnerhas linkeddepersonalization to the processof self- categorization.H ewrites that the factors that enhancethe salienceof in-group– out-group categorizationsenhance the perceivedide ntity between the selfand fellowin- group members;as suc h,the sefac tors ‘‘depersonalize individual self- perception’’ (Turner et al.,1987,p. 50,emphasis in original). In consequence, ‘‘the depersonalizationof self-perception isthe basicproce ss underlyinggroup phenomena’’ (p. 50). Depersonalization,being the basicprocess of group processes,is linke dto arangeof group phenomena.I n-group cohesion, in-group co-operation, socialstere otypingand so on aresaid to be consequencesof depersonalization.Turner writes that ‘‘sub- theoriesof, for example,group cohesion,social co- operation and influencesneedto be developed consistent with existinge mpiricalge neralizationsbut alsoproduc tive of distinctive,testablepredictions and clearly derived fromthe depersonalization process’’ (Turner et al.,1987,p. 56,emphasis adde d). Theimportanc eof the notion of depersonalizationcan alsobe seeninother recenttheoriesderivedfrom SocialI dentity Theory,suchasPostmes, Spears,and Lea(19 99),whose SIDEmodel(the Social Identity Modelof Deindividuationeffec ts) reinterprets deindividuationin terms of depersonalization. Itisnot the present intention to offer acritiqueof thesere cent approaches to .C learlythat iswe llbe yond the scope of the present analysis. Rather the intention ismuc hmorelimited. I tisto point out how the term ‘depersonalization’, inmoving to the theoreticalc entre of laterdeve lopments, has beenaltered intwo crucialrespects. (a) Fromout-group to in-group .Inthe quotations givenabove ,itis c learthat Tajfel, indiscussing depe rsonalization,was writing about the depersonalizationof out-group members—not the depersonalizationof the selfor fellowin- group members.B ycon- trast, the linkageof depersonalizationto the selfand to the in-group isc entralto Self-CategorizationT heory.AsHogg(1996 )writes,‘ ‘self-categorizationdepe rsonalizes self-perception interms of the evaluativelypositive ingroup prototy pe’’(p. 72). (b) Fromnegative topositive .The shift from out-group to the self(and, thus, to the in-group) involvesa shift invalue. I nthe quotations from Tajfel,de personalization,like prejudice,is clearly see nasa problem.I tisthe start of aprocessthat canle ad to dehumanizationand massslaughter. In self-categorizationthe ory,the emphasisis ve ry different. Writers take careto stress the positiveaspe cts of depersonalization.Turne r writesthat depersonalizationshould not be seenasa lossof individualidentity ,rather than as ‘‘a gain inide ntity’’(Turner et al.,1987,p. 51,emphasisin original). A sTurner (1984)stre ssed,‘ ‘depersonalizationin this analysisshould not be consideredaloss(as inthe conceptof de-individuation)but achange,and perhaps evena gain,in identity , asthe individualassumes the characteristicsof aculturallyc onstructedidentity’’ (pp. 535–536 ). Theterminologyof ‘loss’and ‘gain’is significant, in c onstruing ‘depersonalization’as somethingthat canpositive lybenefit the individualby providinga socialidentity . Depersonalizationis not assumedto be asocialproblem. Thus itis to be distinguished from terms suchas‘ dehumanization’and ‘deindividuation’that havety pically appearedin social psy chologicaltheorizing as social problems (see,most notably, Zimbardo,1969). Thus, H ogg(1996) write s: Cognitiveaspects of prejudice 183

Nothing negativeis implied by the term ‘depersonalization’. Itcontains none of the implicationsof ‘dehumanization’or ‘deindividuation’, but simplyrefers to a contextual change inthe levelof identity,not to alossof identity (p. 69,emphasis inoriginal). Insuch comments the double shift inthe treatment of ‘depersonalization’is apparent. Thefoc us has been transformedfrom ‘prejudice’againstout- groups to the positive benefit to the in-group member.The possible continuum betweendepersonalization and dehumanizationis not giventheore ticalspac e.Infact, the linkageis theore tically denied,for ‘depersonalization’is said to carry no implication of ‘dehumanization’ Tajfel’s ideathat depersonalizationmight on occasionsbe the first step to dehumaniz- ation isomitted inthis shift of theoreticalattention. The double shift alsoappe arsas T urner (1999)appe arsto dismisspre judiceand stereotypingas fundamental problems.T urner writesthat depersonalizationis the basicproce ss that ‘‘producesgroup behaviour’’ and adds ‘‘psychologicalgroup for- mation isan adaptiveproc ess’’ (p. 14).Thisperspec tive,he claims,provide sanew wayof thinkingabout prejudiceand stereotyping.I nthis new wayof thinking,‘ preju- dice’ceasesto be aproblemassuc h,at leastin the way that itis traditionally conceived.T he negativeconnotation of ‘stereotype’is re jected.Stere otypesarenot to be seenasinvalid or inaccurate.‘‘Stereotypicac centuation reflects the rationalse lec- tivityof perception inwhich it is more appropriate to seepeoplein some c ontexts at the levelof socialc ategoryidentity than at the levelof personalide ntity’’(p. 26).T hus stereotypingdoes not ‘‘impoverish,but enrichessocialperc eption’’ (p. 27).I ndeed, part of self-categorizationrese archis devote dto exploringthe extent to whichstereo- typingmight be ‘‘veridical’’ (p. 28;se eStangor,1995, for criticismof researchthat attempts to seestereotypesasaccurate or veridical). The problematicof ‘Cognitiveaspe cts’has shifted. Tajfelin that articlestarte dfrom the assumption that prejudiceand stereotypingwere social problems. H ewasc on- cerned to tracetheirorigins to rationalfac tors. Indoingthis, he wasnot arguingthat prejudiceand stereotypingwere the mselvesrational, but that irrationalconsequenc es did not needto be explainedin te rmsof irrationalmotivations. I nTurner’s Self- CategorizationTheory ,prejudicedoes not appearas a probleminits own right.N or doesthe continuum between depersonalizationand dehumanizationfeature .Iffol- lowedrigidly ,this perspectivewould makean analysisof the Holocaust evenmore inappropriate.O newould not wish to imaginee xtremebigotry asan‘enriching’social perception, nor to seethe Nazistere otypesof Jewsbeing describe dasrational,‘ appro- priate’or ‘adaptive’bec ausethe yaresoc iallyrather than individuallybase d.E venthe ideaof exploringwhether or not such stereotypingmight be ‘veridical’is distasteful. Whateverthe merits of the Self-CategorizationTheory —and this isnot to deny that the approach mightnot haveconside rableforc einsome c ontexts—it isaddre ssedto a different setof problems.

Towardsa studyof bigotry Thereis not spacehereto do anythingmore than makesuggestions about recastingthe socialpsy chologicalstudy of bigotry.Suchrecastingcan take, as its starting point, the theoreticalinte rsticesof ‘Cognitiveaspe cts’. Severalfac tors would needto be taken into account to developthe continuum betweenprejudiceand bigotry or between depersonalizationand dehumanization.Some pre liminarypoints canbe madeabout suchadevelopment. 184 MichaelBillig

(i) Bigotry as ideological .Inlinewith Tajfel’s argumentin ‘ Cognitiveaspe cts’, itc an be reaffirmedthat the study of bigotry should not be reducedto the personaldy namics of individualbigots. B igotryty picallyis more than anindividualemotional inve stment. Itisa feature of group relations.I nthis regard,the ideologicalbasisof bigotry needs to be recognized,as inde ed Tajfelacknowledgedwith hisconc ept of assimilation.

(ii) Discursive basis ofideology .Thedisc ursiveposition insocial psy chology stressesthe keyroleof languagein the socialworld (i.e.B illig,1987a; E dwards, 1997; Potter&Wetherell,1987). I deologiesareabove all discursive ,instantiated within discursiveactions (Billig,1991). T hus, the categoriesof ideology,together with shared stereotypingand commonplacesociale xplanations,are frame dinlanguage .Ifthis isac cepted,it is no longernec essary to understand categorizationin te rmsof models that arede rivedfrom perceptual processes,as is commonly found inmost cognitive socialpsy chology,includingTajfel’ s ‘Cognitiveaspec ts’. Instead, categorizationand stereotypingc anbe investigatedwithin discursiveinterac tion.

(iii) Emotionalaspects ofideology .Theide ologyof bigotry cannot be seenmerely asacognitiveappraisal of socialreality .Thelimitations of apurelyc ognitiveapproac h to intergroup relationshave be enrecognizedby mainstreamsocial psy chologists.F or instance,Stephan and Stephan (1985)emphasize the importanceof ‘intergroup anxiety’,but this primarilyrefers to individualcontact acrossgroups rather than the ideologyof hatred(see alsoM aas& Schaller,1991). B rewer(1999)argue sfor examin- inghatred. However,she offers no characterizationof hatredinhe rargumentthat in-group preferenceprovides‘ ‘aplatform for outgroup hate’’(p. 434).She states that thereis a ‘‘fineline be tweenthe absenceof trust and the presenceof activedistrust’ ’ (p. 435)but doesnot indicatethe nature of that dividingline. What isre quired isnot merelythe addition of ‘emotionalvariable s’to the prevailingc ognitivepe rspectivebut atheoreticalre assessment of the apparent distinction betweencognitionand emotion. Ifideologiesare said to encompass emotions, this does not mean that emotions should be seen asfree -floatingpsy chologicalimpulse s,ly ingbe hind ideologiesor socialc at- egories.The yexist within sociallyshared explanations,blamings, ac countings and so on. The emotion within anideologyof hatred isnot somethinge xtra that isadde dto a cognitiveinte rpretation; itis part of that interpretation. WhenTajfelwaswriting, itwas customary for psychologiststo assumea rigidsplit be tween cognitionand emotion, or betweenrationaland irrationalaspe cts of human functioning. The social constructionist position, which has beendevelopedsubse quently,attempts to bridge this gap.T hus, socialc onstructionists havestressed the socialand discursiveconstitu- tion of emotions (Billig,1999; E dwards, 1997;Harreé&Gillett, 1994).I nconsequence, to say that bigotry involvesemotions does not implythat theremust be an emotional force behind the bigotry.Theemotions willbe contained within the hate-talkthat comprisesthe bigotry.

(iv) Reconceptualizing depersonalization and dehumanization in discursive terms.FollowingT ajfel,one canassert that the notions of depersonalizationand dehumanizationare vital to an understandingof bigotry.Depersonalizationshould not be restrictedto the depersonalizationof the self,nor should itbe seenasa cognitive processthat somehow liesbehind language.I nstead, the focus should be on the ways that particularway sof speakingmight de personalizethe ‘other’. Toprobe this further, one would want to examinethe languageof stereotypingas use dinac tualsocial interaction. Dehumanizationwill be an extremeform of depersonalization,as the Cognitiveaspects of prejudice 185

‘other’is de pictedassome how less than human. Dehumanizationmay ,for instance, occur ine thnic jokes,and itwill ce rtainlybe found inthe extremesof hate-talk(B illig, 2001).The ideology of bigotry,inwhich the discourseof dehumanizationwill occ ur, mightbe presumed to be an emotion-ladendiscourse. (v) Repressed and unrepressed emotions .Tajfel,in c ommon with most cognitive socialpsy chologists,was resolutely anti-Freudian.H edid not wish the study of ideol- ogyto be reducedto individualor interpersonaldy namics.H owever,a discursive approach offers the possibilityof reconstituting Freudianthe ory around the notion of repression ina way that avoidsindividual reductionism (Billig,1997, 1999). This perspectiveassume sthat languageis both expressiveand repressive:in order to speak appropriately,speakersmust learnto repress routinelythe desireto speak inappropri- ately.Inany socialconte xt, therewill be norms and routines that permitc ertain discursiveactions and that forbid others. What issocially forbidden can becomean object of desireand pleasure.I fthere aretaboos on the expression of bigotry in contemporary society,outward prejudicemay take the form of aforbidden pleasure. Bigotry,then,bec omesa temptation. (vi) Pleasure in bigotry .Thiswould lead to the disturbingpossibility that thereis pleasurein bigotry .AsSartrere cognizedinhis Portrait ofthe anti-Semite (1948), the bigotmight enjoy the act of hatred,e speciallyif this includesthe pleasureof doing somethingthat isforbidden. Dehumanizingthe other canbe enjoyableas the bigotis freedfrom the constraints of respect, toleranceand reasonableness.T hus one should not expectextremeracist propaganda to be devoidof humour and mockery (Billig, 2001).The c omplexrelations betwee nlanguageand pleasurein ex tremerac ist humour involve,as Sartre realized, moc kingthe restraints of logicand reason. This would leadto afurther paradoxin the study of prejudice.The‘ Cognitiveaspe cts’de alt with the paradoxof socialirrationality stemming from rationalmotive s.The reis also the possibilitythat the extremesof hatred bringpleasure to the hater.W ereit not so, bigotry would not remainsuch adangerous temptation that constantly needs to be watched. Thesenotions areske tched out inorde rto stress asimplepoint. Aturn towards the study of bigotry and the emotions contained with hate-discoursedoes not meana return to the sort of blood-and-guts psychologythat Tajfelex plicitlyre jected.W hen Tajfelwaswriting, the choiceseemed to be betweenconcentrating on eitherc ognitive or instinctualdy namics,as if the two belongedto entirelyse parated psychological realms.F acedwith that choice,Tajfelunderstandably ,and with good reason, chose the cognitiveaspe cts.Today ,suchastark choiceisunnec essary.The academic,cultural and politicalc limatehas changed.I nmany spheres,there has been ablurringof boundariesthat werepreviously thought to be impermeable.P hilosophically,the firm divisionsbe tween truth and falsityor betweenrationalityand irrationalityno longer seemto fit the current mood: whereasonce academicssought absolute truths and completerationality ,the provisionallyreasonable seems to suffice nowadays.Such shifts havebee nboth exploredby and revealedwithin the writingsof thinkers suchas Richard Rorty,Anthony Giddens and Zygmunt Bauman.Within social psy chology, socialconstruc tionismis a product of this changedmood. More generallythe re mightbe atheoreticalgain if the previouslyac cepteddivisionsbetwee nemotion and cognition,or betweenrationalityand irrationality,no longerhaveto be accepted at leastin the old manner.T he challengeis to explorethe interconnections, especiallyin relationto the continuingproblems of prejudiceand bigotry. 186 MichaelBillig

The politicalc ontextof today should not be forgotten. Thec urrent ageis not one of confident politicalide ologicaltruth. Nevertheless,bigotry remainsa majorsoc ial issuein E urope with fascist partiesappealingto anti-foreignersentime nts, and the re-emergenceof extreme,often violent,nationalist politic s.F ascismin We stern Europe refuses to waneto the point of disappearance,and thereis a markedwax ingin the East.T heseforce sarenot the products of individualmotivations, nor do theyrepresent meresy stems of socialc ategorization.The yrepresent powerful and dangerous mixtures.So long as this ideologicalmixture continuesto threaten,the needto attend to the socialpsy chologicaldy namicsof bigotry persists.

Acknowledgements

The author isgrate fulto CharlesAntaki,D avidM andeland MargaretWethe rellfor the irhe lpful comments on an earlierdraft. P artsof theargume nt of thispape rwerepresente datthespe cial Tajfelsy mposium organizedby theH istoryand Philosophy Sectionof theB ritishP sychological Society,atitsannual conference,held in York, 2000.

References

Billig,M .(1978). Fascists:A socialpsychological view of the National Front .London: Academic Press. Billig,M .(1985).P rejudice,c ategorizationand particularization:F rom aperceptualto a rhetoricalapproac h. EuropeanJournal of SocialPsychology, 15 , 79–103. Billig,M .(1987a). Arguingand thinking .Cambridge:C ambridgeU niversityPress. Billig,M .(1987b).Anti-semiticthemes and theB ritishfar le ft:Some social- psychologicalobser- vationson indirectaspec tsof thec onspiracytradition. In C.F.Graumann &S.Moscovici (Eds.), Changing conceptions of conspiracy (pp. 115–136). N ewYork: Springer. Billig,M .(1991). Ideology and opinions .London: Sage. Billig,M .(1996).Rememberingthe background of socialide ntitytheory .In W.P .Robinson (Ed.), Socialgroups and identities: Developing the legacyof Henri Tajfel . London: Butterworth-Heinemann. Billig,M .(1997).The dialogicunconsc ious:P sychoanalysis,disc ursiveps ychology and the natureof repression. BritishJournal of SocialPsychology, 36 , 139–159. Billig,M .(1999). Freudian repression .Cambridge:C ambridgeU niversityPre ss. Billig,M .(2000). Rock’n rollJews .Nottingham,U K:FiveL eaves. Billig,M .(2001).H umour and hatred:T he racistjoke sof theK uKluxK lan. Discourse and Society, 12, 291–313. Brewer,M .B.(1999).The psychology of prejudice:Ingroup loveor outgroup hate? Journal of SocialIssues, 55 , 429–444. Brown, R.,&Capozza,D .(2000).Socialidentity theory in retrospectand prospect.In D. Capozzaand R.Brown (Eds.), Socialidentity processes .London: Sage. Burke,K .(1962). Agrammarof motives .Cleveland,O H:MeridenB ooks. Cohen, D.(1977). Psychologists on psychology .London: Routledgeand KeganP aul. Edwards,D .(1991).Categoriesare for talking. Theoryand Psychology, 1 , 515–542. Edwards,D .(1997). Discourse and cognition .London: Sage. Eisenberg,L .(1972).The human natureof human nature. Science, 176 , 123–128. Eiser,J. R.(1996).A ccentuation revisited. In W.P .Robinson (Ed.), Socialgroups and identities: Developing the legacyof Henri Tajfel .London: Butterworth-Heinemann. Goldhagen,D .J. (1998). Hitler’s willing executioners .London: Abacus. Harreé,R.,&Gillett,G .(1994). Thediscursive mind .London: Sage. Cognitiveaspects of prejudice 187

Hogg,M .A.(1996).Intragroup proc esses,group structureand socialidentity .In W.P .Robinson (Ed.), Socialgroups and identities: Developing the legacyof Henri Tajfel . London: Butterworth-Heinemann. James,C .L.R.(1964). Beyond aboundary .London: SportsmanB ook Club. Levi,P. (1987). If this is aman .London: Abacus. Lorenz,K .(1940).DurchDomestikationverursachte Sto örungenarteige nen Verhaltens. Zeitschrift fu¨rangewandte Psychologie und Charakterkunde,59 , 2–81. Lorenz,K .(1943).D ieange borenenFormenmoöglicherErfahrung. Zeitschrift fu¨r Tierpsychologie,5 , 235–409. Lorenz,K .(1974). Civilized man’s eight deadly sins .London: Methuen. Lorenz,K .(1976). Behind the mirror .London: Methuen. Maas,A .,&Schaller,M .(1991).Intergroupbiases and thec ognitivedy namicsof stereotype formation. EuropeanReview of SocialPsychology, 2 , 189–209. Mandel,D .R.(1998).T he obediencealibi:M ilgram’s accountof theH olocaustreconsidered. Analyse und Kritik,20 , 74–94. Mandel,D .R.(2001).Instigatorsof genocide:ExaminingH itlerfroma socialpsy chological perspective.In L.S.Newman &R.Erber(E ds.), What socialpsychology can tell us about the Holocaust .Oxford:O xford UniversityPre ss. Nisbett,A .(1976). KonradLorenz .London: J. M.Dent. Postmes,T .,Spears,R .,&Lea,M .(1999).Socialidentity ,normativecontent and ‘deindividua- tion’in computermediatedgroups.In N.Ellemers,R .Spears& B.Doosje( Eds.), Social identity: Context, commitment, content .Oxford:B lackwell. Potter,J., & Wetherell,M .(1987). Discourse and socialpsychology .London: Sage. Robinson, W.P. (Ed.)(1996) . Socialgroups and identities: Developing the legacyof Henri Tajfel.London: Butterworth-Heinemann. Sartre,J. P. (1948). Portrait of the anti-Semite .London: Seckerand Warburg. Stangor,C .(1995).Content and applicationinac curacyin socialste reotyping.In Y.-T.Lee, L.J. Jussim& C.R.McCauley( Eds.), Stereotype accuracy .Washington,D C:American PsychologicalAssociation. Stephan, W.G .,&Stephan,C .W.(1985) .Intergroupanx iety. Journalof SocialIssues, 41 , 157–175. Tajfel,H .(1969a).Cognitiveaspec tsof prejudice. Journalof BiosocialSciences, Supplement No. 1, 173–191. Tajfel,H .(1969b).Soc ialand culturalfac torsin perception. In G.Lindzey &E.Aronson (Eds.), Thehandbook of socialpsychology (2nd ed.,V ol.3, pp. 315–394). R eading,M A:Addison- Wesley. Tajfel,H .(1970).Experimentsin intergroupdiscrimination. Scientific American,223 (5), 96– 102. Tajfel,H .(1981). Human groupsand socialcategories .Cambridge:C ambridgeU niversityP ress. Tajfel,H .(1984).Inte rgroup relations,soc ialmy thsand socialjustice in socialpsy chology.In H.Tajfel(Ed.), Thesocial dimension .Cambridge:C ambridgeU niversityP ress. Tajfel,H .,Billig,M.,Bundy,R.P.,&Flament,C .(1971).Socialcate gorizationand intergroup behaviour. EuropeanJournal of SocialPsychology, 1 , 149–178. Tajfel,H .,&Turner,J. C.(1979).Anintegrativetheory of intergroup conflict.In W.G .Austin& S.Worchel(E ds.), Thesocial psychology of intergroup relations (pp. 33–47). M onterey,CA: Brooks/Cole. Tajfel,H .,&Wilkes,A .L.(1963).C lassificationand quantitativejudge ment. BritishJournal of Psychology, 54 , 101–114. Turner,J. C.(1984).Socialidentific ationand psychologicalgroup formation.I nH.Tajfel(E d.), Thesocial dimension .Cambridge:C ambridgeU niversityPress. Turner,J. C.(1999).Some curre nt issuesin researchon socialidentity and self-categorization theories.In N.Ellemers,R .Spears& B.Doosje( Eds.), Socialidentity: Context, commitment, content.Oxford:B lackwell. 188 MichaelBillig

Turner,J. C.,Hogg,M .A.,Oakes,P .J.,R eicher,S. D .,&Wetherell,M .(1987). Rediscovering the socialgroup .Oxford:B lackwell. Wetherell,M .(1996).Constructingsoc ialidentities: T he individual/socialbinary in HenriT ajfel’s socialpsy chology.In W.P .Robinson (Ed.), Socialgroups and identities: Developing the legacyof Henri Tajfel .London: Butterworth-Heinemann. Whillock, R.K.,&Slayden, D.(1995). Hate speech.Thousand Oakes,C A:Sage. Zimbardo, P. (1969).The human choice:Individuation,reason and orderversusde individuation, impulseand chaos. NebraskaSymposium onMotivation, 17 , 237–307.

Received3 July2000; revised version received24 November 2000