1 
 

 

Race trouble: Attending to race and racism in online interaction 

(2014, British Journal of Social Psychology, DOI: 10.1111/bjso.12070) 

 

 

 

 

 

 

 

 

Kevin Durrheim* 

(University of KwaZulu-Natal) 

 

Ross Greener 

(University of KwaZulu-Natal) 

 

Kevin A. Whitehead 

(University of the Witwatersrand) 

 

 

 

 

 

* Address correspondence to: Kevin Durrheim, Discipline of Psychology, University of KwaZulu-Natal, 

Private Bag, X01, Scottsville, 3209, South Africa. Durrheim@ukzn.ac.za  

mailto:Durrheim@ukzn.ac.za


2 
 

 

Abstract 

This article advocates the concept of Race Trouble as a way of synthesizing variation in racial discourse, 

and as a way of studying how social interaction and institutional life continue to be organized by 

conceptions of “race” and “racism”. Our analysis of an online discussion at a South African University 

about the defensibility of a characterization of (black) student protesters as “savages” revealed a 

number of familiar strategies: participants avoided explicit racism, denied racism, and denied racism on 

behalf of others. However, the aim of analysis was not to identify the “real” racism, but to show how 

race and racism were used in the interaction to develop perspectives on transformation in the 

institution, to produce social division in the University, and to create ambivalently racialized and 

racializing subject positions. We demonstrate how, especially through uses of deracialized discourse, 

participants’ actions were observably shaped by the potential ways in which others could hear “race” 

and “racism”. Race trouble thus became manifest through racial suggestion, allusion, innuendo and 

implication. We conclude with a call to social psychologists to study the ways in which meanings of 

“race” and “racism” are forged and contested in relation to each other.   



3 
 

 

Discursive social psychologists introduced a novel way of studying racism in social psychology. The 

dominant cognitive tradition in the discipline at the time viewed racism as psychological prejudice which 

could be measured with attitude scales. In contrast, discursive researchers adopted a constructionist 

approach, focusing on situated evaluative practices, and analyzed talk and text to show how the 

“objects” of prejudice were variably constructed, and to what rhetorical and ideological ends (Potter & 

Wetherell, 1987; Wetherell & Potter, 1992).  

 

Right from the outset it was abundantly clear that race talk was attuned to the “norm against prejudice” 

(Billig, 1988) and that these interested constructions of outgoups were accompanied by denials of 

racism (van Dijk, 1992). Much discursive research came to focus on how speakers produced this 

ambivalent discourse that expressed beliefs about race while denying racism – as exemplified by the 

disclaimer, “I’m not prejudiced, but…” (Hewitt & Stokes, 1975). Augoustinos and Every (2007) reviewed 

five ways in which negative views about outgroups could be presented as being reasonable and justified: 

(1) denying prejudice; (2) portraying views as reasonable and rational reflections of the external world; 

(3) developing positive self-presentations – e.g., as being tolerant and hospitable – in contrast to 

negative other presentations, e.g., as criminal, deviant, and culturally alien; (4) discursively deracializing 

negative representations of outgroups using non-racial terms; and (5) using liberal arguments about 

freedom, equality and progress as justifications.  

 

This research highlighted the action-oriented and rhetorical nature of racial attitudes (see Billig, 1987). 

That is, speakers produce (potentially) race-relevant positions in the service of particular actions or 

arguments, rather than simply as an expression of cognitively based views or beliefs, and they do so in 

ways that anticipate possible hearings of their talk as racist or prejudiced (Whitehead, 2009). Race 

attitudes are thus not the product of individual minds in isolation, but are produced in interaction 



4 
 

 

through practices that reveal how potential responses have shaped their formulation. Condor et al. 

(2006) argue that such race talk “represents an essentially collaborative exercise” as speakers anticipate 

how their and other people’s utterances may be heard, and then preempt discrediting responses (p. 

458).  

 

This interactional character of race attitudes suggests that constructions of race are supported by 

contrasting constructions of racism. In order to deny prejudice, Billig (1999) argues, speakers need to 

develop a serviceable account of “the ‘proper’ prejudice” that is to be denied (p. 152). “Racism”, then, is 

as much a variable object of construction as “race” is. For example, arguments that present racial 

attitudes as reasonable and empirically-grounded are implicitly underpinned by the view that prejudice 

is irrational and ignorant. On the other hand, portrayals of racial attitudes as understandable fears can 

be excused if prejudice is considered an unwarranted emotional reaction motivated by intolerance 

(Figgou & Condor, 2006). Alternatively, if racism is defined as prejudgment, racial attitudes can be 

presented as beliefs reluctantly arrived at (Edwards, 2003). Or, in a strategy that van Dijk (1992) called 

reversal, dominant group members can position themselves as the real victims of racism.  

  

More recently, discursive researchers have begun to study a countervailing norm, not against prejudice, 

but against accusations of prejudice and racism (Augoustinos & Every, 2010; Goodman & Burke, 2010, 

2012). This research demonstrates how accusations of racism can be criticized for illegitimately “playing 

the race card” and attempting to shut down reasonable debate. This offensive rhetoric does the same 

thing as the defensive rhetoric of racism denial: it constructs an understanding of racism that allows 

speakers to portray their views as not racist. In consequence, critics of racism carefully construct their 

criticisms to undermine the potential accusation that they are playing the race card. Their delicate, 



5 
 

 

euphemistic, and indirect accusations of racism reveal an implicit analysis of what illegitimate 

accusations of racism might involve. 

 

Ideas about racism thus play a complex and multifaceted role in structuring debates about potentially 

race-relevant matters. The direction and outcome of these debates depend on successful constructions 

of racism, where “the success, or otherwise, of a claim to non-prejudiced character ultimately depends 

upon its acceptance or rejection on the part of an audience” (Condor et al, 2006, p. 548). Contending 

versions of racism are produced when people talk about race explicitly or implicitly, and when they 

counter explicit or potential accusations of racism or of having played the race card. The meaning of 

racism is up for grabs in such contestation, and is defined in the end by whatever is ratified as racism, 

and the policies and practices that stem from this. 

 

In this article we suggest that the animating role that “racism” plays in public debate and controversies 

may be described and conceptualized as “race trouble” (Durrheim, Mtose & Brown, 2011). This concept 

draws attention to four features of the struggle over the nature of racism that we have considered.  

1. “Race” and “racism” are co-constitutive. The nature of racism is constructed in serviceable ways 

to enable the expression or criticism of particular, situationally circumscribed, racial attitudes.  

2. “Racism” is produced in interaction. The meaning of racism is constructed in interaction, 

dialogue and debate. It is forged in the space between speaking and hearing; and care needs to 

be taken to see how racial utterances are constructed to anticipate potential hearings and are 

responded to.  

3. “Racism” is often implicit. A primary method of attending to potential accusations of racism is to 

employ deracialized discourse (Augousinos & Every, 2007). The logical endpoint of this strategy 

is to eliminate all explicit references to race and to invoke race indirectly by allusion and 



6 
 

 

innuendo (Whitehead, 2009; Durrheim, 2012; Bergseiker et al., 2012). Ideas about race and 

racism may thus be produced in the minds of the hearers (and taken into account) without 

explicit reference in the words of the speakers. However, as soon as race and racism become 

topics of debate, race trouble moves from the realm of implicit shared knowledge to explicitly 

racialized discourse.  

4. “Racism” is institutional. Racism isn’t solely the province of individual racists, or racist 

discourses. It is also an institutionally-located collaborative accomplishment that functions as 

part of a “racial formation” (Omi & Winant, 1994). This draws attention to the temporal and 

situated processes by which race and racism are constructed and put to work, both in 

developing social policies, institutions and agendas, and at the level of everyday experience and 

interaction.  

 

The analysis of race trouble thus draws attention to the structuring of interaction around particular 

issues; how “race” and “racism” organize social, psychological and institutional life; the practical and 

material consequences of this organization; and the way these change over time. Consequently, the 

concept of race trouble helps to lay the groundwork for a critical social psychology focused on the 

analysis of interaction rather than the definition and identification of racism. We follow the lead of 

Condor et al., (2006) in understanding “racism” as a “collaborative accomplishment”, the “responsibility 

for which is shared jointly between the person of the speaker, and those other co-present individuals 

who occasion, reinforce or simply fail to suppress it” (p. 459). We are thus cautious about restricting our 

focus to “racist discourse” because (1) this may foreground analysts’ definitions of racism, whereas (2) 

what counts as racism is a collective production in which all participants – radical, liberal and 

conservative – rely on conventional wisdom about race and racism, and (3) the interaction might not be 

explicitly about race or racism, but only implicitly so. 



7 
 

 

 

The remainder of the article provides a demonstration of race trouble at work in an online discussion 

forum – called the “Change listserve” – which was established for employees at a South African 

university to discuss transformation in that institution. Like all universities in South Africa, this university 

had been struggling with transformation in the aftermath of apartheid. It is a historically white 

institution, which has undergone successive waves of transformation in leadership, policies, 

organizational structure and staff and student demographics. The student body changed most 

dramatically as the all-white apartheid university now has a majority of black students. Change in the 

composition of staff has been much slower to implement, but the university has promoted affirmative 

action aggressively, and has sought to root out all forms of white privilege. Needless to say, all this 

change has not been without controversy and concerns about race and racism have always been near 

the surface, if not explicitly articulated. Analysis of race trouble directs attention to studying the 

institutional alignments, mobilizations and the organization of social life in these debates, showing how 

race trouble works there, and to what effect.  

 

The case study 

The Change listserve was established in June 2006 with the stated aim of providing “a virtual meeting 

place for those who are interested in improving the climate of communication” on campus. We 

downloaded the complete database of posts as at the end of March 2011. It consisted of a total of 2396 

posts arranged in 1008 conversation threads. The discussion forum and the community it represented 

were defined early in its life, in a vitriolic debate between a core group of founding members, the 

university principal, and a leading figure of the Black African Academic Forum (BAAF). The BAAF was an 

informal group of academics who had produced a policy document containing wide-ranging proposals to 

challenge racism and to Africanize the University. A number of core members of the Change listserve 



8 
 

 

had met to discuss the vision of transformation published by the BAAF. They were especially “concerned 

about the increasing racialization of the debate on transformation” (thread number 8; post 1, 

21/June/2006), and later developed a petition informed by principles of “non-racialism, academic 

freedom and integrity; and good governance” (thread 46, post 1, 08/August/2006). These views of 

transformation were dismissed by a member of the BAAF as being “not rooted in the expereince of 

Aficans”, and as being representative of people who have “taken sides with the priviledged and have 

iginred the plight of the underpriviledged” (thread 44, post 8, 06/August/2006)1. 

 

These fundamentally divergent views of transformation were perhaps best expressed when a BAAF 

member challenged the taken-for-granted dominance of English in the University: “We want to have 

African dreams and values back into this institution you wish to transform. Next time I will write you and 

your friends in isiXhosa mfowethu [my brother], my father tounge, or even isiZulu, and articulate my 

views for an institution of African scholarship I wish to see” (thread 44, post 13, 15/August/2006). The 

BAAF proposal to accept isiZulu as an official language would certainly Africanize the institution and 

make learning more accessible to mother tongue isiZulu speakers, but it would also exclude the majority 

of the traditional staff and student body who were unable to speak an African language. These views 

were greeted with incredulity by many members of the Change listserve, one of whom argued that “The 

racism of one era has been exchanged for the racism of another, a development we have for too long 

been pretending is not happening. Who's willing to fight this as we fought apartheid?” (thread 44, post 

14,  16/August/2006).  

 

This early debate served to define the Change community in a number of ways. They were positioned as 

non- or anti-African and as resisting transformation. Although not explicitly stated, this could easily be 

read as positioning the Change community as a racist “white old guard”.  Of course, the members of the 



9 
 

 

change listserve did not see themselves in this way. Many positioned themselves on the political left and 

prided themselves for having opposed apartheid. They claimed that the BAAF and its proposal was 

racially exclusive, exclusionary, and hence racist. Moreover, many felt that their jobs were threatened 

by their being positioned as opposing the Africanization agenda of the new University management, and 

they were concerned that their comments on the listserve were being monitored. 

 

It is thus evident that racism was at stake in the debate, with different versions being pitted against each 

other in this struggle for the soul of the institution. It is tempting to identify the “real” racism, perhaps 

applying modern racism theory to show how one grouping was subtly trading in symbolically racist 

sentiment or imagery. Or perhaps their discourse could be analysed to show how it developed an 

interested version of race and transformation and sought to deny racism. Such analyses would amount 

to taking sides and thus taking part in the debate about transformation, advancing the kinds of 

arguments that the participants themselves were producing, deploying ideas about racism to formulate 

a politics, economics, or pragmatics. The study of race trouble, by contrast, is an analysis of how ideas 

about race and racism are used to organize the interaction under investigation, and the personal and 

collective investments in it.  

 

Methods 

The analysis focuses on two discussion threads that dealt with the violent nature of student protests. By 

treating each thread as a sequence and each posting as a turn, we were able to analyse this 

asynchronous discussion by applying the methods of conversation analysis (Sacks, Schegloff and 

Jefferson 1974; Schegloff 2007) and discursive psychology (Edwards & Potter 1992; Edwards 2005). We 

were particularly interested to see how the participants managed matters of stake in the discussion, 

which was implicitly and explicitly about race and racism. To this end, we sought to describe how shared 



10 
 

 

knowledge of race and racism was developed and treated as such in an interactional context where 

“hearers must perform active contextualizing work in order to see what descriptions mean, and 

speakers rely upon hearers performing such work in order that their utterances will make definite 

sense”  (Heritage, 1984, p. 148; cf. Edwards, 2004). As our analysis shows, race trouble is manifest as 

participants in the discussion (1) used shared resources of race and racism as tools for explaining and 

accounting; and (2) addressed or avoided addressing these uses of race and racism as topics of 

discussion. 

 

I can’t understand them… the savages 

Extract 1 begins early (post 4) in a newly established thread (number 777) entitled “Protests”. The 

participant uses a nom-de-plume to protect their true identity in a post which conveys very strong 

opinions about student protesters (cf. Billig, 1989). As a new member of the listserve the participant has 

a relatively high poster number (197), and this contribution received critical response from a 

longstanding member of the listserve, Poster 27. 

 

Extract 1 

Post 4; Poster 197, 23/March/2009 17:00 

1.  If I call the protesters a bunch of savages will I be protected by academic 

2.  freedom? 

3.  If they have a legitimate protest then they have a duty to let us know what it 

4.  is about. The violence directed at those in no position to help them is 

5.  sickening and indicative of a larger malaise in the protest culture so evident 

6.  in South Africa.  They are eroding their support base by attacking fellow 

7.  students and staff that may give a shit.  I do admit that with their behavior I  

8.  am all for excluding them for life.  Maybe a protest should start with a  

9.  statement of 

10. 1) what is the issue? 



11 
 

 

11. 2) what is their expected outcome? 

12. 3) What is the time frame for results? 

13. 4) Give a chance for follow up and redress 

14. 5) etc 

15. As opposed to lets trash the joint and go from there.  By the time we find out  

16. what they want we won't care. 

 

Post 5; Poster 27, 23/March/2009 17:12 

1.  Last week I suggested, in jest, that someone could use a pseudonym and write  

2.  some much needed satire on Change. However, I actually feel that it  

3.  is not a good idea if people post their opinion on a range of issues using a  

4.  pseudonym.  

5.  We should assert our right to freedom of speech, and own our ideas as much as  

6.  is possible. I realize that some people don't feel free to express their  

7.  views. And yet, I think it's not good for the exchange of ideas on Change 

8.  if we don't state who we are. I'd be interested to know what others  

9.  think.  

 

Post 6; Poster 197; 23/March/2009 17:24 

1.  Well my last post would probably get me assaulted and accused of all sorts so  

2.  I am all for a nom-de-plume. 

 

In post 4, Poster 197 (P197) criticizes the student protesters for behaviour that is claimed to be violent, 

wrongly directed (“at those in no position to help them”) and impulsive (“lets trash the joint and go from 

there”). This portrayal of violent, impulsive and inarticulate protesters underscores their 

characterization as “a bunch of savages”, and is highlighted by contrast with a more civilized and 

reasoned approach suggested by the five-part list in lines 10-15. Although there are no explicit 

references to race, there is evidence that race is an issue. Most obviously, the word savage is a 

powerfully loaded racial “curse word” (Bakhtin, 1981) in post-colonial Africa, especially when it is used 



12 
 

 

to depict actions of black people as impulsive, violent and stupid. Other Change members reading the 

post would certainly know that the student protesters were black. In this period of transformation, 

student protest at the University was always undertaken by black (mostly financially disadvantaged) 

students who experienced many challenges including material deprivation, financial exclusions, and 

inadequate housing. In this regard, the student protest was similar to other, increasingly common, 

“service delivery protests” by poor and under-resourced black communities – a link made explicit in 

P197’s description of the protest as “indicative of a larger malaise of protest culture so evident in South 

Africa”.  

 

Not only is the racial identity of the protesting “savages” implicitly available, but so too is the identity of 

P197. These kinds of depictions are hearable as South African “white talk” (Steyn, 2001). The expression 

is framed in the indirect and delicate manner associated with the denial of racism. Despite the strong 

nature of his/her opinions (including expletives and extreme recommendations), P197 disavows the use 

of the word savages, by marking it as hypothetical with the structure, “If I...will I…” P197 also displays a 

reflexive recognition that the expression may breach a taboo by marking it as a formulation that might 

need protection by academic freedom. But the reasons for this are unstated and remain open to 

interpretation in non-racial terms – including the alternative association of savages with crowd 

members (cf. Reicher, 2001).  

 

None of these hearable concerns about race and racism are directly attended to in the immediately 

subsequent posts. Instead, P27 debates the value of using a pseudonym, distancing him/herself from 

P197 by claiming that an earlier suggestion to use a pseudonym was made in jest. The absence of 

explicitly racial discourse, however, does not mean the absence of race trouble. The potential hearing of 

racism must be actively avoided. P27 acknowledges that some Change members “don’t feel free to 



13 
 

 

express their views” on “a range of issues”, thereby displaying an understanding P197’s reason for using 

a pseudonym, but skirts reference to “savages”, thereby avoiding a direct response to the possible 

racism of post 4. The deracialized criticism of P197 is met with a deracialized defense in post 6. By 

speculating that responses to the earlier post might include assault and accusations, P197 acknowledges 

that the post was objectionable, but skirts the reference to savages and hearable racism as possible 

reasons for this. 

 

The way the participants politely skirt around the topic of race and racism suggests that these themes 

are being dialogically repressed by collaborative avoidance that can “look as if there is a joint conspiracy 

to achieve a collective refusal of knowledgeability” (Billig, 1997, p. 152). At the same time, racial themes 

are potentially invoked by the references to savages, views that people are not free to express, and 

violent reactions to these views. Such gestures allow hearers to apply racial common sense themselves 

to arrive at a fuller meaning of what is being said. They enable hearers to quietly use racial stereotypes 

and categories to understand what is happening in the interaction and the University without their 

having to be avowed in florid, impolite and unspeakable detail (Durrheim, 2012; Whitehead, 2009). The 

possibility that racial innuendo renders the post problematic is confirmed later in the discussion (see 

Extract 3), where accusations of racism become explicit. In the meantime, racism is the proverbial 

“elephant in the room”. As we will show, posters who avoid engaging directly with this potential 

interpretation expose themselves to criticism of colluding with racism by not naming it. This shows how 

the commonsense norm against making accusations of racism is matched by the opposing norm against 

racism, and that all participation in the discussion is dilemmatic (cf. Billig, 1987). 

 

Extract 2 is a continuation of this discussion thread, omitting two posts that avoided referring to race 

and “savages” but offered opinions about academic freedom and the use of nom-de-plumes. In post 9, 



14 
 

 

P30 indirectly opposes the use of pseudonyms, arguing that participants should be accountable for 

posting “insulting and inflammatory” material such as P197’s “indefensible” comment about “students 

as ‘savages’”. These remarks are then followed by a number of posts in which the hearable racism of the 

reference to the protesters as savages becomes progressively more directly topicalized.  

 

Extract 2 

Post 9; Poster 30; 23/March/2009 19:12 

1.  And yet, I think that if we are in fact accountable for what we say, we will  

2.  be much more likely to evaluate whether it is important, helpful and worthy of  

3.  debate, or if it is insulting and inflammatory. 

4.  ((Poster 197)), I think that your comment about students as 'savages' is  

5.  indefensible, even if it may be your 'right' to express it. I can think of 

6.  many reasons to avoid such statements other than fear of assault or 

7.  disciplinary action against you.  

 

Post 10; Poster 168; 24/March/2009 08:46 

1.  Thank you [Poster 30] for this response to ((Poster 197)). I absolutely support  

2.  an argument against pseudonyms if it reduces the probability of the change  

3.  list being used as a vehicle for self protection from such indefensible 

4.  prejudicial idiocy of such comments as those made by ((Poster 197)). Change 

5.  has been established to provide a forum for debate borne out of a  

6.  motivation for responsible participation for a just institution and society.  

7.  Could we please respect it as such and not use it as a platform for bigotry and  

8.  prejudice rooted in historical oppression under the fallacious guise of the  

9.  'right to' freedom of speech.  

 

Post 11; Poster 199; 24/March/2009 09:27 

[11 lines omitted that contain “self-reflection” about the meaning of “responsible 

participation”] 

12. ((Poster 197)), you provide some useful pointers to protest strategy. Which  



15 
 

 

13. were lost on people (and would be so on the students) as a result of your  

14. opening line. Perhaps you were on the receiving end of or witnessed  

15. people/property being damaged etc. and your reaction stemmed from that. Like  

16. you I don't want to see SA going down the twin slippery slopes of oppressive  

17. unaccountable government and reactionary destructive responses. But when  

18. conflicts erupt we especially need to keep our cool, to think before reacting,  

19. etc. That's also when the relationships we've established with others provide a  

20. foundation (or reveal the lack thereof) for being able to respond  

21. constructively.  

 

Post 12; Poster 27; 24/March/2009 09:58 

[17 lines omitted, the poster gives a general update on the protest action on campus] 

 

Post 13; Poster 4; 24/March/2009 13:15 

[7 lines omitted, the poster recommends compulsory module 'Rights and Obligations of 

the SA Citizen'] 

 

Post 14; Poster 24; 24/March/2009 15:39 

1.  I do wonder whether referring to a word, 'savages' (plus some other choice  

2.  phrases) to conclude that this is '..indefensible...' and  '... a platform for  

3.  bigotry and prejudice rooted in historical oppression  under the fallacious  

4.  guise of the 'right to' freedom of speech', isn't jumping to conclusions and  

5.  making a lot of assumptions about an author's intent and premises. Such  

6.  responses can inadvertently be a form of censorship by castigation -that only  

7.  polite, non-contentious, nice language is allowed.  Its the sort of response  

8.  that is the despair of satirists  

[6 lines omitted, the author of the post repeats their position stated above] 

 

P30’s criticism of P137 continues to be deracialized in post 9. Although this post characterizes the 

comment about students as savages as “indefensible”, it does not explicitly identify the grounds of 

indefensibility. Rather, P30 invites readers to imagine these by saying that s/he “can think of many 



16 
 

 

reasons to avoid such statements” besides the “fear of assault or disciplinary action” that had been 

suggested previously by P197. Race is thus being alluded to by inviting readers to consider why 

references to savages might be “indefensible” (cf. Whitehead, 2009). In the following post, P168 agrees 

with this evaluation, upgrading the censure to “indefensible prejudicial idiocy”, and arguing that the 

listserve should not be used as a “platform for bigotry and prejudice rooted in historical oppression”. It 

is further noteworthy that P168 prefaces his/her post by thanking P30 for his/her response to P197, 

which suggests an orientation to the intervening posters having failed to take the opportunity to 

similarly rebuke P197, and thus having missed the “elephant in the room”. 

 

Although these two forthright rebukes provide the clearest treatment yet of P197’s post as racialized, 

they nevertheless avoid direct reference to race, and avoid accusing P197 of racism. These concerns are 

strongly alluded to but race and racism remain implicit elements of the background. Even so, posts 11 

and 14 provide defenses of P197 that have all the hallmarks of a denial of racism made on behalf of 

someone else (Condor et al., 2006). Post 11 takes the form of a disclaimer. P199 concedes that the 

opening line of the original post was unfortunate but supplies a reason why someone might say 

something like this: P197 may have been “on the receiving end of or witnessed people/property being 

damaged”. The original “indefensible” characterization of the students as “savages” is thereby reduced 

to a “reaction” rather than a habitual or enduring prejudice indicative of racism – having “an axe to 

grind” (Figgou & Condor, 2006) – and is thereby treated as being, in fact, defensible. P199 thus provides 

an alternative explanation to the inference of racism, which has not been explicitly made but which is 

now treated as requiring denial.  

 

Similarly, in post 14, P24 criticizes the critics for “jumping to conclusions and making a lot of 

assumptions about…the intent and premises” of P197. This argument that the critics are prejudiced – 



17 
 

 

making judgments without evidence – follows the pattern of denial of racism by reversal (van Dijk, 

1992), but avoids directly referring to racism by leaving the nature of the unsubstantiated judgments 

unspecified. P14 orients to the norm against making accusations of racism (Augoustinos & Every, 2010) 

both by criticizing the critics for making prejudicial judgments and by arguing that such criticism “can 

inadvertently be a form of censorship by castigation” (cf. Goodman, 2010; Goodman & Burke, 2012). 

Despite the fact that the offending post made no explicitly racist utterance, by this point in the thread 

the accusations, denials, and counter-accusations strongly suggest that racism is the issue. They allude 

to race and racism and take recognizable form as denials of racism. Nonetheless, they are deniably 

about race and racism because these themes have not been explicitly mentioned and direct accusations 

of racism have not been made. This changes in Extract 3 when the posters explicitly treat the previous 

discussion as having been about race and racism. 

 

Extract 3 

Post 15; Poster 197; 25/March/2009 09:47 

1.  Admittedly some of my words were not great due to a historical usage of a  

2.  certain word.  My words were born out of frustration at violence and threats of 

3.  violence as a means of protest in the first instance and in no way refer to any 

4.  population group so assumed. 

[17 lines omitted, the author provides multiple dictionary definitions of the word 

‘savage’] 

23. And I will use a nom de plume due to violence and the threat of disciplinary  

24. action. 

 

Post 16; Poster 168; 25/March/2009 11:06 

1.  ja, ja...and in a perfect world where all people are free and equal (which we  

2.  know this world of ours is not) all words can have equal weight.  I don't  

3.  think we should muffle censure of oppression through over-determining 

4.  political correctness (which is after all concerning power (im)balances). Open 



18 
 

 

5.  debate on terminology, and issues like race and racism should not, in the name 

6.  of freedom of expression, become an unwitting platform for condonation of 

7.  racist proclamations and practices - which by definition are 'silencing' 

8.  oppressive mechanisms. In this case, the word 'savage' is a racially loaded 

9.  term in our context. When used in reference to black students it is racist - 

10. pure and simple. If my response silences racism - frankly that is a good 

11. thing.  ...and yet...the point is taken in regard to the untamed words used 

12. in critical response...but too much taming...? 

 

Post 17; Poster 198; 25/March/2009 12:16 

1.  ((Poster 168)) and others,   

2.  I won't defend the use of the term - savage, but I will not condemn it either.  

3.  One may agree that the term 'savage' is historically loaded in the African  

4.  context, but is not the responses to its use also historically loaded?  ie - 

5.  the assumption being made is that the students were indeed 'black'?  And that 

6.  the author meant the term in such a manner?  Indeed no reference was made to 

7.  the race of the students, just their behaviour.   

8.  In fact I read (maybe even misread) the statement as tongue in cheek about the  

9.  protests and issues of academic freedom rolled into a simple, cheeky question.   

10. What it drew my attention to (even if unwittingly) was that the protesters 

11. were largely comprised of 'black' students thus raising questions of real 

12. race issues of access and equality at ((the University)).  And makes us 

13. wonder to what extent does academic freedom go?  Should we not have open 

14. discussions about topics that make us uncomfortable?  Make us angry?   

 

The first explicit mention of race in the thread occurs in post 15, as P197 re-enters the discussion to add 

voice to the denial that has become progressively more clearly about race. Even here, however, the 

problem is attributed to hearers who have made unwarranted assumptions; and race is referred to 

obliquely by way of the euphemism, “population group”, a term that has sometimes been used in place 

of “race” in surveys and in the South African census. P197 aligns with the argument in posts 11 and 14 



19 
 

 

that the “indefensible” comment about students as “savages” was a momentary reaction, “born out of 

frustration at violence and threats of violence”, not other motives and intentions. The racial nature of 

the disavowed alternative motive for use of the word savage is suggested by P197’s admission that the 

problem with the expression related to the “historical usage of a certain word” and the denial that the 

utterance referred to “any population group”. Although no explicit accusations of racism have been 

made by this point, and post 15 is not an explicit denial of racism, this admission and explanation further 

suggest that concerns about race and racism have been hearably present and are now being treated as 

requiring explicit denial. 

 

Race and racism then become explicit in the final two posts. Post 16 develops an explicit accusation of 

racism, saying that referring to black students as “savage” is “racist – pure and simple”. P168 also 

criticizes the previous 11 posts by arguing that political correctness and debates about terminology 

might muffle censure and lead to an unwitting “condonation of racist proclamations and practices”. The 

previous posters are thus sanctioned for not naming racism when it was hearably there, which provides 

strong retrospective evidence that there was hearable racism present previously in the thread, and that 

the lack of explicit responses to it are treatable as an accountable absence. On the other hand, because 

race and racism have not been explicit, P168 is open to the change of over-hearing race and of 

developing an unwarranted accusation of racism. This possibility is taken up in the final post when P198 

responds to P168, noting that “no reference was made to the race of the students”, thereby charging 

P168 with racializing the discussion or “playing the race card” (Whitehead, 2012; Capdevila & Callaghan, 

2008) by making the assumption that the word applied specifically to black students. P168’s response is 

thus problematized for being “historically loaded”, based on (ideological) assumptions and not on facts. 

P198 further develops this denial of racism on behalf of P197 by saying that s/he read the comment as 

“tongue-in-cheek”, lacking racist intent.  



20 
 

 

 

In addition to denying that the original reference to students as “savages” lacked racial intent, P198 

argues that the hearable racial allusion might indeed have anti-racist effects. P198 claims that the 

“offending” post drew his/her attention (albeit “unwittingly”) to the demographic character of the 

protesting students, which in turn had raised “questions of real race issues of access and equality” at the 

institution. P198 directs attention to the “real race issues” in the university, suggesting that the 

argument over the use of the word “savages” is a red herring. In the process P198 pre-emptively 

undermines potential accusations of racism or racial insensitivity to which s/he could become vulnerable 

as a result of defending a poster who has just been accused of racism (cf. Whitehead, 2013).    

 

Discussion 

The analysis has shown how posters avoided making explicitly racialized statements, denied racism, and 

denied racism on behalf of others. All these strategies are well documented in the discursive literature 

and reveal the dilemmatic and ambivalent way in which people orient to the norm against prejudice at 

the same time as articulating and defending beliefs that are hearably about race (Billig, 1988; Billig et al., 

1988). Participants also generally avoided making direct accusations of racism and criticized accusations 

of racism for shutting down debate and limiting academic freedom. They were thus also oriented to the 

norm against making accusations of racism (Goodman, 2010; Goodman & Burke, 2012).  

 

The most striking feature of this discourse was the absence, until late in the discussion, of explicit 

references to race and racism. This mode of expression is strategically important because deracialized 

discourse allows speakers to deny racism (Augoustinos & Every, 2007), but also to deny that they have 

made accusations of racism. At the same time, carefully formulated deracialized discourse can be quietly 

heard to be about race. Such hearings are evident, for example, when someone denies racism on behalf 



21 
 

 

of another (Condor et al., 2006), when racism is denied in the absence of explicit accusations (e.g., 

Whitehead & Wittig, 2004); or when speakers produce what appear to be “designedly ambiguous”, and 

thus deniable, utterances (Stokoe, 2012a, p. 282) – i.e., descriptions and actions that may imply 

particular categories and associated inferential upshots, but which appear to be designed to be 

ambiguous, such that they cannot be definitively identified as category-relevant (also see Stokoe, 

2012b). These hearings, rather than topicalized racism, were the main object of concern as the 

participants grappled with the defensibility of the representation of the student protesters as savages. 

This expression was suggestive of colonial discourse and was thus vulnerable to being heard as indexing 

race indirectly. The immediately subsequent criticism portrayed the statement as indefensible but 

avoided stating why. These deracialized accusations of hearable racism were met by deracialized denials 

that offered alternative explanations of the motives underlying the expression and the intentions of the 

poster. Further deracialized counteraccusations suggested that the critics were prejudiced (and 

undermining freedom of expression) but avoided specifying the grounds of their prejudice.  

 

All this indirection and deracialized discourse suggests that the posters were oriented to norms against 

prejudice and accusations of racism. Since race and racism were not explicit, they were readily deniable, 

and any critic could be blamed for hearing racism in the absence of explicit talk about race. Critics used 

the same tactic, making deracialized accusations to avoid the counter-accusation that they were playing 

the race card. Finally, all these refusals to properly hear racism – by avoiding these topics altogether 

(post 10), or not by not explicitly naming racism (post 16) – are subject to the criticism that they are 

colluding with and condoning racism; and this criticism is itself subject to critique as the value of anti-

racism is opposed to the value of freedom of speech or some other commonplace.  

 



22 
 

 

The deracialized discourse studied here made these hearings of race and racism possible because it 

gestured toward these themes obliquely by suggestion, implication, allusion or innuendo. Racial themes 

were not omitted entirely but were subtly used to provide explanations and accounts. Other research 

has shown that this can be done by posing a puzzle that enables hearers to provide a racial explanation 

(Whitehead, 2009), by “stereotyping by implication” (Durrheim 2012), or even by conspicuously 

omitting a negative stereotype or by allowing “audiences [to] draw negative inferences from 

communicators’ faint or unidimensional praise of targets” (Bergsieker, Leslie, Constantine and Fiske, 

2012, p. 19). Ambiguously racialized words and expressions are the best for accomplishing this. Terms 

like ‘savage’ and ‘population group’ are not explicitly about race and are easy to defend as having 

nothing to do with race (cf. Stokoe, 2012a; 2012b). At the same time they allude to race, providing 

innuendo or suggestion that the right audience would be able to hear.  

 

Are we confronted here with the problem of racism and racist discourse? As analysts, should we join 

P168 in condemning the original post for racism and the subsequent posters for guilt by silence? There 

are reasons for caution. First, all participants positioned themselves against racism. They generally also 

positioned themselves as being in favor of freedom of expression and displayed caution in making 

accusations of racism. Moreover, each positioning in the debate was criticizable in terms of race and 

racism as the very meaning of these constructs and their appropriate use was being debated. The 

meaning of “racism” and the identity of the “racists” cannot be nailed down easily as accusations are 

met by counter-accusations in the unfolding debate. Diagnoses of racism overshadow these participant 

maneuverings with analysts’ authoritative judgments.  

 

For these reasons, Durrheim et al. (2011) recommend analyzing race trouble rather than identifying 

racist discourse. The shift to race trouble is not a way to better understand the dynamics of racism in 



23 
 

 

conversation. Instead it seeks to re-specify the problem: it is about understanding contestations about 

what counts as racism, and understanding the ways in which race "troubles" people. Race trouble is a 

synthesizing concept that provides a way of conceptualizing the entirety of the kind of interaction we 

have considered in this article. Analysis of race trouble synthesizes variation in talk (Potter & Wetherell, 

1987), but the aim is not simply to use variation to study the occasioned and interested nature of 

utterances. Rather, it aims to show how multiple utterances rely on each other by alignment (with or 

against) and by anticipating hearings. The aim is to show how they collectively construct silences or 

implicit meanings which, in turn, can serve as resources for action.  

 

Durrheim et al. (2011) considered various ways in which social life in South Africa was organized by ideas 

about race and racism, defining race trouble as “dynamic mutually constitutive practices and contexts of 

social division” (p. 199). The exchange we analyzed here exemplifies a kind of interaction that is 

premised upon and helps to reproduce racial understanding of the world and one’s position within it. In 

particular, it articulated a view of student protest from the perspective of a community of academics. In 

the process the participants deployed understandings of race and racism to position themselves as 

liberal, radical or anti-racist, at the same time as entertaining racial representations to render the 

current situation intelligible and to develop a perspective on transformation that opposed the radical 

action of the protesting students. The interaction we studied was a facet of broader deliberations and 

interactions about transformation at the University, and it thus contributed to the production of the 

fractured and fractious social context, seeking change but haunted by the racial divisions of the past.  

 

Race trouble of this kind is largely immune to criticisms of racism because racism is already dead – in  

the sense that few openly support it – and the possibility of being accused of racism has already been 

taken into account. In this sense, racism is “undecidable” (Derrida, as cited in Collins & Mayblin, 2000). It 



24 
 

 

is horrific: like a zombie, it cannot killed because it is already dead; but it can’t be ignored either, 

because it remains potent. Attempts to resolve the undecidability are susceptible to become infected by 

it.2  

 

It is important to note, however, race trouble need not be governed by the norms against prejudice and 

accusations of prejudice, such as we have studied here. Explicit – even crude – references to race and 

racism can also help to constitute “dynamic mutually constitutive practices and contexts of social 

division”, as is evident in comments on online newspaper forums (see Cresswell, Whitehead & 

Durrheim, 2014). In these acrimonious and hateful exchanges, participants engage in “interactional 

scaffolding” (Condor & Figgou, 2012) as they accuse each other of and deny racism, and as they escalate 

racial insults until they terminate with or without de-escalation. Here too ideas about race and racism 

are reworked in strategic projects that build alliances and create divisions that reflect and reproduce the 

racial divisions and inequalities of the past.  

 

The meanings of race and racism are thus forged co-constitutively with reference to each other in talk 

and action that is either explicitly or implicitly about race. The concept of race trouble helps us to focus 

on this interaction as a whole rather than targeting a particular subset of arguments and positions 

within it. This means recognizing the role that “anti-racist” discourse can play in anchoring collective 

meanings and action; and how it can become infected by “race” and “racism”. Such analysis seeks to 

understand the historical and intuitional edifice of race trouble; its constitutive positionings and 

alignments; how it shifts and changes over time; its impetus for inertia; and how it serves to sustain 

unequal institutions that are proliferating in the supposedly post-racial world. 

 

Note. 



25 
 

 

1. The spelling and grammatical errors that appeared in the Change discussion are reproduced in 

direct quotations from this data in this article. 

2.  For example, the debate about immigration in Australia and Europe epitomizes a form of race 

trouble in which arguments against immigration are readily heard as being racist but they deny racism, 

taking this hearing into account, and instantiate the norm against accusation of racism by suggesting 

that criticism undermines freedom of expression (Augoustinos & Every, 2010; Goodman, 2010; 

Goodman & Burke, 2012; Burke & Goodman, 2012). Not only do critics of anti-immigration discourse 

have to contend with all this ambiguity and possible sanction for targeting racism, but they can also 

employ the same racial representations of foreigners as their counter-accusers. For example, arguments 

that promote sympathy for refugees fleeing persecution elsewhere are bolstered by a view of the racial 

elsewhere as a violent place that abuses human rights (cf. Every & Augoustinos, 2008). 

 

 

 

References 

Augoustinos, M., & Every, D. (2007). The language of ‘‘race’’ and prejudice: A discourse of denial, 

reason, and liberal-practical politics. Journal of Language and Social Psychology, 26, 123–141. 

http://dx.doi.org/10.1177/0261927X07300075 

 

Augoustinos, M., & Every, D. (2010). Accusations and denials of racism: Managing moral accountability 

in public discourse. Discourse & Society, 21, 251–256. http://dx.doi.org/10.1177/0957926509360650 

 

Bakhtin, M. (1981). The dialogical imagination. Texas: University of Texas Press. 

 

http://dx.doi.org/10.1177/0261927X07300075
http://dx.doi.org/10.1177/0957926509360650


26 
 

 

Bergsieker, H.B., Leslie, L. M., Constantine, V. S., & Fiske, S. T. (2012). Stereotyping by omission: 

Eliminate the negative, accentuate the positive. Journal of Personality and Social Psychology.  

http://dx.doi.org/10.1037/a0027717 

 

Billig, M. (1987). Arguing and Thinking: A Rhetorical Approach to Social Psychology. Cambridge: 

Cambridge University Press. 

 

Billig, M. (1988). The notion of ‘prejudice’: Some rhetorical and ideological aspects. Text, 8, 91-111. 

http://dx.doi.org/10.1515/text.1.1988.8.1-2.91 

 

Billig, M. (1989). The argumentative nature of holding strong views: A case study. European Journal of 

Social Psychology, 19, 203-223. http://dx.doi.org/10.1002/ejsp.2420190303 

 

Billig, M. (1997). The dialogic unconscious: Psychoanalysis, discursive psychology and the nature of 

repression. British Journal of Social Psychology, 36, 139-159. http://dx.doi.org/10.1111/j.2044-

8309.1997.tb01124.x 

 

Billig, M. (1999). Freudian repression: Conversation creating the unconscious. Cambridge: Cambridge 

University Press. 

 

Billig, M., Condor, S., Edwards, D., Gane, M., Middleston, D., & Radley, A. (1988). Ideological Dilemmas. 

London: Sage. 

 

Burke, S., & Goodman, S. (2012). ‘Bring back Hitler’s gas chambers’: Asylum seeking, Nazis and Facebook 

– a discursive analysis. Discourse & Society, 23, 19–33. http://dx.doi.org/10.1177/0957926511431036 

http://dx.doi.org/10.1037/a0027717
http://dx.doi.org/10.1515/text.1.1988.8.1-2.91
http://dx.doi.org/10.1002/ejsp.2420190303
http://dx.doi.org/10.1111/j.2044-8309.1997.tb01124.x
http://dx.doi.org/10.1111/j.2044-8309.1997.tb01124.x
http://dx.doi.org/10.1177/0957926511431036


27 
 

 

 

Capdevila, R, & Callaghan, J. (2008). ‘It’s not racist, it’s common sense’: A critical analysis of political 

discourse around asylum and immigration in the UK. Journal of Community and Applied Social 

Psychology, 18, 1–16. http://dx.doi.org/10.1002/casp.904 

 

Collins, J. & Mayblin, B. (2000). Introducing Derrida. St. Leonards, NSW: Allen & Unwin. 

 

Condor, S., & Figgou, L. (2012). . In J. Dixon & M. Levine (Eds). Rethinking the prejudice problematic: A 

collaborative cognition approach. Beyond prejudice: extending the social psychology of conflict, 

inequality and social change (pp. 200-222). Cambridge: Cambridge University Press. 

 

Condor, S., Figgou, L., Abell, J., Gibson, S., & Stevenson, C. (2006). ‘They’re not racist…’ Prejudice denial, 

mitigation and suppression in dialogue. British Journal of Social Psychology, 45, 441-462. 

http://dx.doi.org/10.1348/014466605X66817 

 

Cresswell, C., Whitehead, K., & Durrheim, K. (2014). The anatomy of ‘race trouble’ in online interactions. 

Ethnic and Racial Studies. DOI: 10.1080/01419870.2013.854920. 

 

Durrheim, K. (2012). Implicit prejudice in mind and interaction. In J. Dixon & M. Levine (Eds), Beyond 

prejudice: extending the social psychology of conflict, inequality and social change (pp. 179-199). 

Cambridge: Cambridge University Press.  

 

Durrheim, K., Mtose, X., & Brown, L. (2011). Race trouble: Race, identity and inequality in post-apartheid 

South Africa. Scottsville: University of KwaZulu-Natal Press. 

http://dx.doi.org/10.1002/casp.904
http://dx.doi.org/10.1348/014466605X66817


28 
 

 

 

Edwards, D. (2003). Analysing racial discourse: The discursive psychology of mind-world relationships, In 

H.van den Berg, H., M. Wetherell & Houtkoop-Steenstra (Eds.), Analysing race talk: Multidisciplinary 

approaches to the interview. Cambridge: CUP. 

 

Edwards, D. (2004). Shared knowledge as a performative category in conversation. Rivista di 

Psicololinguitica Applicata, 4, 41-53. 

 

Edwards, D. (2005). Discursive Psychology, in K.L. Fitch & R.E. Saunders. E. (eds), Handbook of Language 

and Social Interaction (pp. 257-273). Mahwah, NJ: Erlbaum. 

 

Edwards, D. & Potter, J. (1992). Discursive psychology. London: Sage Publishers. 

 

Every, D. & Augoustinos, M. (2008). ‘Taking advantage’ or fleeing persecution? Opposing accounts of 

asylum seeking. Journal of Sociolinguistics, 12, 648–667. http://dx.doi.org/10.1111/j.1467-

9841.2008.00386.x 

 

Figgou, L., & Condor, S. (2006). Irrational categorization, natural intolerance and reasonable 

discrimination: Lay representations of prejudice and racism. British Journal of Social Psychology, 45, 

219–243. http://dx.doi.org/10.1348/014466605X40770 

 

Goodman, S. (2010). ‘‘It’s not racist to impose limits on immigration’’: Constructing the boundaries of 

racism in the asylum and immigration debate. Critical Approaches to Discourse Analysis across 

Disciplines, 4, 1–17. 

http://dx.doi.org/10.1111/j.1467-9841.2008.00386.x
http://dx.doi.org/10.1111/j.1467-9841.2008.00386.x
http://dx.doi.org/10.1348/014466605X40770


29 
 

 

 

Goodman, S and Burke, S (2010) “Oh you don’t want asylum seekers, oh you’re just racist”: A discursive 

analysis of discussions about whether it’s racist to oppose asylum seeking. Discourse and Society, 21, 

325–340. http://dx.doi.org/10.1177/0957926509360743 

 

Heritage, J. (1984). Garfinkel and ethnomethodology. Cambridge: Polity Press. 

 

Hewitt, J. P., & Stokes, R. (1975). Disclaimers. American Sociological Review, 40, 1–11. 

http://dx.doi.org/10.2307/2094442 

 

Omi, M., & Winant, H. (1994). Racial formation in the United States: From the 1960s to the 1990s 

(revised edn). New York: Routledge. 

 

Potter, J., & Wetherell, M. (1987). Discourse and social psychology: Beyond attitudes and behaviour. 

London: Sage. 

  

Reicher, S. D. (2001). The psychology of crowd dynamics. In M. A. Hogg & R. S. Tindale  (Eds.), Blackwell 

Handbook of social psychology: Group processes (pp. 182–208). Oxford, UK: Blackwell. 

 

Sacks, H., Schegloff, E. A. & Jefferson, G. (1974). A simplest systematics for the organization of turn 

taking in conversation. Language, 50, 696-735. http://dx.doi.org/10.2307/412243 

 

Schegloff, E. (2007). Sequence organization in interaction: A primer in conversation analysis. Cambridge: 

Cambridge University Press. 

http://dx.doi.org/10.1177/0957926509360743
http://dx.doi.org/10.2307/2094442
http://dx.doi.org/10.2307/412243


30 
 

 

 

Steyn, M. (2001). Whiteness just isn't what it used to be: White identity in a changing South Africa. 

Albany: State University of New York Press. 

 

Stokoe, L. (2012a). Moving forward with membership categorization analysis: Methods for systematic 

analysis. Discourse Studies, 14(3), 277-303. http://dx.doi.org/10.1177/1461445612441534 

 

Stokoe, L. (2012b). Categorial systematics. Discourse Studies, 14(3), 345-354. 

http://dx.doi.org/10.1177/1461445612441534 

 

van Dijk, T. A. (1992). Discourse and the denial of racism. Discourse & Society, 3, 87-118. 

http://dx.doi.org/10.1177/0957926592003001005 

 

Wetherell, M., & Potter, J. (1992). Mapping the language of racism: Discourse and the legitimation of 

exploitation. Hertfordshire: Harvester Wheatsheaf. 

 

Whitehead, K. A. (2009). Categorizing the Categorizer: The Management of Racial Common Sense in 

Interaction. Social Psychology Quarterly, 72, 325-342. http://dx.doi.org/10.1177/019027250907200406 

 

Whitehead, K. A. (2012). Racial Categories as Resources and Constraints in Everyday 

Interactions: Implications for Racialism and Non-Racialism in Post-Apartheid South Africa. Ethnic and 

Racial Studies, 35, 1248-1265. http://dx.doi.org/10.1080/01419870.2011.591407 

 

http://dx.doi.org/10.1177/1461445612441534
http://dx.doi.org/10.1177/1461445612441534
http://dx.doi.org/10.1177/0957926592003001005
http://dx.doi.org/10.1177/019027250907200406
http://dx.doi.org/10.1080/01419870.2011.591407


31 
 

 

Whitehead, K. A. (2013). Managing Self/Other Relations in Complaint Sequences: The use of Self-

Deprecating and Affiliative Racial Categorizations. Research on Language and Social Interaction, 46, 186-

203. http://dx.doi.org/10.1080/08351813.2013.780342 

 

Whitehead, K. A., & Wittig, M. A. (2004). Discursive management of resistance to a multicultural 

education curriculum. Qualitative Research in Psychology, 1(4), 267-284. 

 

http://dx.doi.org/10.1080/08351813.2013.780342