Constructing racial hierarchies of skill – Experiencing affirmative action in a South African organisation : A qualitative review

Democratic South Africa is characteristic of massive social and economic inequalities, which are based largely on racial lines (Durrheim, Mtose & Brown, 2011; Franchi, 2003), most of which stem from the apartheid era. Whilst it is acknowledged that society is becoming desegregated and inequality deracialised, the extent to which racial transformation is actually occurring is debateable. And, whilst attempts to iron out historic injustices are being made, past patterns of inequality and segregation continue to feature and, at the same time, ‘new patterns have emerged that continue to be structured around race’ (Durrheim et al., 2011, p. 21). In an attempt to address historic injustices, the South African government has made concerted efforts to ‘deracialise’ South Africa politically, economically and socially, most notably through the use of affirmative action (AA) measures within the labour market. The rationale behind this is that, since the implementation of the Employment Equity Act (1998), AA provides a platform from which to change the demographic weighting of disadvantage in the workplace.


Introduction
Democratic South Africa is characteristic of massive social and economic inequalities, which are based largely on racial lines (Durrheim, Mtose & Brown, 2011;Franchi, 2003), most of which stem from the apartheid era.Whilst it is acknowledged that society is becoming desegregated and inequality deracialised, the extent to which racial transformation is actually occurring is debateable.And, whilst attempts to iron out historic injustices are being made, past patterns of inequality and segregation continue to feature and, at the same time, 'new patterns have emerged that continue to be structured around race' (Durrheim et al., 2011, p. 21).In an attempt to address historic injustices, the South African government has made concerted efforts to 'deracialise' South Africa politically, economically and socially, most notably through the use of affirmative action (AA) measures within the labour market.The rationale behind this is that, since the implementation of the Employment Equity Act (1998), AA provides a platform from which to change the demographic weighting of disadvantage in the workplace.
Efforts, in the form of AA, to integrate South Africa's historically disadvantaged into the labour force have been met with practical and ideological barriers from all areas of society.Given these barriers, the current research is important for at least four reasons.
Firstly, there is relative lack of research that qualitatively examines AA in South Africa, and even less from within the discursive tradition (Durrheim, Boettiger, Essack, Maarschalk & Ranchod, 2007;Durrheim & Dixon, 2005).Furthermore, much of the research in this area tends to be one-sided in that it is largely focused on the perceptions of white South Africans.There is a pervasiveness of negativity surrounding AA.This negativity, despite being around for many years, is still a fervently contested and controversial subject.Kennedy-Dubourdieu (2006, p. 2) states that AA is a social policy that 'engenders an inflamed debate and opinion polls consistently reveal that practically everyone has an opinion on the subject, even though there is a great deal of confusion over what the policy actually entails'.For many employees, AA is seen only as a compromise that, in itself, perpetuates the discrimination it seeks to address.Expressions of fear, racial tension and discrimination still surface and, as Romano (2007) suggests, whilst many South Africans claim to be in favour of AA, AA policies still generate considerable amounts of criticism.
Given the varied experiences of AA, this research sought to give expression to people's subjective realities on the issue by looking at how everyday practices function contextually to give meaning to social and psychological life, specifically related to the ways in which we frame and conceptualise our experiences of AA.This study thus aimed to explore AA from a social constructionist orientation with a focus on Potter and Wetherell's discursive psychology.
The language that we use every day is a 'dynamic form of social practice which constructs the social world, individual selves and identity' (Potter, 1996, p. 118).Given this, a second reason for this research was to examine how the way in which people use language can function in perpetuating historic privilege for some, and historic disadvantage for others.
Given South Africa's unique socio-political terrain, there is a need to critically engage with AA from a perspective with which to locate forms of 'meaning' within the broader social and cultural context that informs subjective realities (Burr, 1995;Edwards, 1997;Edwards & Potter, 1992;Potter & Wetherell, 1987).To date, only a few researchers have approached AA from this perspective (Augustinos et al., 2005;Duncan, 2003;Franchi, 2003;Kravitz et al., 1996;Stevens, 2003).
A third reason for the current research is that most AArelated research from the discursive tradition seems to draw from related studies of race and race relations, with fewer studies looking at AA specifically (Augustinos & Every, 2007;Durrheim & Dixon, 2005;Van Dijk, 1993).
Fourthly, and equally important is to consider, is that most studies in this area, regardless of the theoretical and methodological approaches adopted, consider AA in polarised terms: people either support or oppose AA in the abstract (Reyna, Tucker, Korfmacher & Henry, 2005).The current study was interested in the more complex picture, particularly in relation to what people think about the policy conceptually as opposed to exclusively looking at their more practical experience of AA.In other words, this research endeavours to contribute new knowledge, specifically from the South African context, regarding the value, or lack thereof, associated with the principles embedded in AA policies.

Research purpose
The main purpose of this study was to explore employee's social constructions of AA in a South African organisation.
The current study used social constructionism as a theoretical framework, with a focus on Potter and Wetherell's discursive psychology to explore how everyday practices function contextually to give meaning to social and psychological life, specifically related to the ways in which we frame and conceptualise our experiences of AA.Broadly speaking, social constructionism is a theoretical orientation that offers critical alternatives in psychology such that understanding, knowledge and meaning are located within a context (Burr, 1995).Potter and Wetherell's discursive psychology is embedded within the social constructionist paradigm and involves a radical rethinking of concepts with a focus on psychological themes (Edwards & Potter, 1992;Potter & Wetherell, 1987).
The current research sought to achieve the following two aims: 1.The first aim was to critically analyse the discourses around AA produced by employees within in a racially diverse, privately owned South African organisation within the retail sector. 2. The second aim was to explore how the historically advantaged and disadvantaged employees in this organisation construct AA.
In achieving these aims, the discursive resources and rhetorical arguments that employees used in talking about their perceptions and experiences of AA were investigated.

Contribution of the study
There are few documented studies in South Africa that adopt social constructionism as an approach using diverse samples.Thus, the current study has shown that it is useful to adopt a social constructionist approach within the South African context, particularly when questioning taken-for-granted knowledge and trying to understand phenomena that are historically and socially specific.The study's methodology highlights the real value in studying context in meaningmaking and in studying implied inferences that underlie talk.
The findings may contribute to the discipline of industrial psychology, particularly in the ways in which policies around preferential treatment are conceptualised, and subsequently implemented, within organisations.This knowledge may contribute to the ways in which organisational policy documents are conceptualised so that attempts may be taken to move beyond the things that serve to perpetuate inequality within the workplace.
Furthermore, the findings also highlight the possibility for other organisations to consider that, despite their impressive policies of inclusion and transformation, there may exist a disjuncture between the intensions of the organisation on the one hand, and the experiences of employees on the other.

Literature review
By way of introduction, AA is first discussed in general terms.The review then moves on to highlight the need for a more subjective, contextualised approach (in the form social constructionism) to study how people frame and contextualise AA.A brief description of social constructionism, as the study's theoretical framework, is provided.The literature review ends with the presentation of related studies that used social constructionism, with particular focus on the seminal work of Potter and Wetherell (1987).

Affirmative action
Affirmative action is described as a: range of governmental and private initiatives that offer preferential treatment to members of designated racial or ethnic groups … usually as a means of compensating them for the effects of past and present discrimination.(Swain, 1996, p. 1) In principal, these initiatives seem fair, especially given the historic discrimination previously faced by black South Africans.In practice, however, AA is particularly complex, often presenting conflicting opinions on whether the policy is fair or not (Bobbitt-Zeher, 2013;Charlton & Van Niekerk, 1994;Gloppen, 1997;Human, Bluen & Davies, 1999, Jordan-Zachery, 2012;Kelbaugh, 2003;Mazzocco, Cooper & Mariagrace, 2012;Tummala, 1999).Fuelling this debate is the distinct polarisation of attitudes toward AA practices, which arguably originate from historical deprivation, political ideology and a sense of personal and collective loss (Franchi, 2003).AA is still a fervently contested and controversial subject.Importantly, Bentley and Habib (2008) reflect the thoughts of many South Africans when they state that South Africa's democracy is faced with the political dilemma of how to address historical injustices whilst, at the same time, being able to build a single national identity and promote economic growth and development.Expressions of fear, racial tension and discrimination are equally met with expressions of democracy, freedom and equalising opportunities in South African organisations.Romano (2007) suggests that whilst many South Africans claim to be in favour of AA, AA policies still generate considerable amounts of criticism.As a result, efforts to integrate the country's historically disadvantaged into the labour force have been met with practical and ideological barriers from all areas of society.
Importantly, this ambivalent opposition to AA suggests the salience of exploring the extent to which employees embrace AA in the workplace, especially given that South Africa is a relatively new democracy with much of its inclusive policies still in their infancy.

From realism to social constructionism
Mainstream psychology, and by default realism, largely views behaviour as being influenced by an objective truth and internal mental states (Parker, 1990).Mainstream psychology, being essentialist and realist in nature, proclaimed the existence of an 'essential' core within people that can be identified and explained (Gough & McFadden, 2001).It accepts that people have an essential, inherent, identifiable nature.Realist approaches seek to understand pre-existing attitudinal functions and structures (Potter & Wetherell, 1987) in the absence of contextual specificity.Durrheim and Dixon (2005, p. 448) describe realist approaches as impoverished realism in that the world is 'stripped of its particularity and nuance'.Most of the literature on AA in South Africa is realist in nature (Adam, 2000;Charlton & Van Niekerk, 1994;Cohen & Sterba, 2003;Gloppen, 1997;Human et al., 1999;Kelbaugh, 2003;Sachs, 1992;Sono & Werner, 2004;Tummala, 1999), with little research considering how people themselves frame and conceptualise AA.As a result, knowledge about AA may only shed limited light on the multiple, shifting meanings that may be attached to AA.Using non-realist methodologies not only highlights the different meanings attached to AA but, more importantly, it allows one to critically engage with AA in South Africa in a way that is appreciative of its rich socio-political history.Against this backdrop, the current study adopted social constructionism as an approach to study the subjective, contextualised experiences of AA and thus moves away from the realist perspective.

Social constructionism
Noted earlier, an important interest in the current study was to critically engage with the embodied nature of prejudice (Ainsworth, 2000) that stems from everyday practices, particularly in reference to the ways in which we conceptualise and experience AA.
In response to this, many researchers have taken the 'discursive' turn to focus on research that is qualitative in nature.As a result, there has been recently a strong movement towards understanding AA within social psychology, and more specifically from the tradition of discursive psychology, an approach conceptualised by Potter and Wetherell using social constructionism.A social constructionist approach views language as a dynamic form of social practice which gives expression to subjective psychological realities.
Social constructionism was first made popular in 1966 with Peter.L. Berger and Thomas Luckmann's book, The social construction of reality.The authors argued that everything we know, including taken-for-granted knowledge and common sense, is created and sustained through social practices and social interaction.Put simply, social constructionism advocates that when we use language, or discourses, we actually create the world in which we live, and 'that the very nature of ourselves as people, our thoughts, feelings and experiences, are all the result of language' (Burr, 1995, p. 33).Thus, the way in which we, for example, may talk about AA can function in rationalising existing social inequalities and, at the same time, in denying prejudice (Augustinos, Tuffin & Every, 2005;Duncan, 2003;Franchi, 2003;Kravitz, Harrison & Turner, 1996;Stevens, 2003).Language, therefore, is much more than a form of communication: it is also a system of representation.Thus, by employing social constructionism as an approach to understanding people's experiences, we can study how everyday practices, such as 'talk', can function contextually to give meaning to social and psychological life, specifically related to the ways in which we frame and conceptualise our experiences of AA.Against this reasoning, the current study used discourse as a method of enquiry to critically engage with articulated productions of AA in South Africa in a way that is appreciative of its rich socio-political history.
Interestingly, there have only been a few previous studies specifically on AA that have been undertaken in the discursive tradition, both internationally and in South Africa (e.g.Augustinos et al., 2005;Franchi, 2003 Potter andWetherell, 1987;Wetherell & Potter, 1992).Most research focuses on race and race relations, prejudice and, amongst other things, modern racism -all largely focused on identifying the pervasive repertoires and devices that are used by people to justify social inequalities (e.g.Augustinos & Every, 2007;Duncan, 2003Durrheim et al., 2007;Durrheim & Dixon, 2005;Franchi, 2003;Potter, 1996;Potter & Wetherell, 1987;Van Dijk, 1993;Wetherell, Stiven & Potter, 1987;Wetherell, Taylor & Yates, 2001).Apart from the content, however, these related studies are particularly useful in that they illustrate how language use performs social actions that construct varying versions of the world (Potter & Wetherell, 1987).
For example, Wetherell et al.'s (1987) study, which focused on university students' talk related to employment opportunity and gender, is one of the earlier, and very influential, studies within the discursive paradigm.The authors were interested in exploring student views of the status of employment opportunities for women.They were specifically interested in studying the ideologies surrounding the reproduction of gender inequality.The study revealed contradictory repertoires at play, which functioned in justifying existing gender inequalities in both the work and home sphere.
In particular, their findings illustrated a conflict between students on the one hand supporting equal opportunities and on the other hand emphasising the factors supposedly limiting those opportunities.These limiting factors were presented as practical constraints; for example, participants drew comments about how women bearing children and subsequently rearing children posed challenges to their progression.Importantly Wetherell et al. demonstrated how this type of talk constructed 'unequal egalitarianism', which both appreciates equality and at the same time justifies the limitations in not achieving it.Social change, pertaining specifically to the women's ability to progress, was seen as the responsibility of the woman and in her ability to prove her equivalence (Wetherell et al., 1987).
In their study, Potter and Wetherell (1987) found that Pakeha (white New Zealanders) often legitimised their opposition to AA measures for the Maori (native New Zealanders) by drawing on discourses of meritocracy and 'togetherness'.The meritocratic discourse functioned in portraying AA as problematic in that it defied the principles of meritocracy where individuals should be rewarded based on merit.Secondly, the togetherness discourse functioned in portraying AA as destructive in that preferential treatment could result in disharmony amongst those who benefit and those who do not benefit from preferential treatment.On the whole, the study showed how participants constructed AA as problematic because it did not adhere to the principles of justice and fairness.Later, in 1992, the authors (Potter and Wetherell) went on to study an analysis of 'race', again with New Zealanders.Expanding on their earlier study, the authors found that Pakeha participants drew on a range of egalitarian principles (such as fairness and freedom) in an attempt to justify existing unequal social relations between the two groups.The authors highlight that these discourses were presented as being rhetorically self-sufficient.For example, some statements made were 'everybody should be treated equally' and 'everyone can succeed if they try hard enough' (Wetherell & Potter, 1992, p. 177).
In a study of a similar nature, Augustinos et al. ( 2005) conducted a study amongst Australian undergraduate students in an attempt to study race relations in Australia.Their analysis built on previous studies in the discursive tradition on AA with a particular interest in illustrating how participants drew on resources to construct AA as largely problematic.In doing so, their research presented a discursive analysis of conversations produced from focus groups discussions on race, disadvantage and AA.The findings of the study suggest that opposition to AA tended to be justified by liberal, egalitarian principles and self-sufficient arguments such as 'everyone should be treated equally'.Furthermore, the authors also found a meritocratic discourse at play which identified merit as being most important regarding entry into tertiary education.The study also showed how participants' talk was constructed and put together in a manner that 'presented' speakers as fair and reasonable.The authors discuss how contradictory discourses are reflective of competing values and how the language of the 'new racism' is framed by ideological dilemmas and ambivalence (Augustinos et al., 2005).
As noted earlier, there is limited research on AA that adopts a social constructionist perspective.However, apart from content, related studies are particularly useful in that they illustrate how language use performs social actions that construct varying versions of the world.In this case, a case study of symbolic racism (Franchi, 2003) is used as an illustration.Franchi (2003) critically analyses the discourses produced by 33 employees in a training workshop that was designed to address issues of racialised conflict and to promote intercultural sensitivity.The findings revealed that race continues to feature in the ways to process information about the 'self' and 'other'.The study also suggested that whilst South Africans have changed their language on race, their assumptions about 'racial symmetry' still feature, albeit in more subtle forms.Furthermore, the findings highlight that participants who opposed AA generally constructed AA beneficiaries as 'inferior' in a way that functioned to legitimise the maintenance of white participants' status and power.
Franchi also found the use of temporal markers such as 'now' and 'in the older days' functioned in delegitimising AA practices, which, by default, emphasised the 'legitimacy' of the past.In both instances, AA is constructed as troublesome which inadvertently presented the historic status quo as justifiable.
These studies showcase the ways in which historic privilege and unequal power relations continue to manifest in society through everyday language use.Furthermore, this also illustrates that the discursive tradition is useful to gauge what people think about the policy conceptually as opposed to exclusively looking at their more practical experience of AA.

Research approach
This study aimed to explore AA from a social constructionist orientation with a focus on Potter and Wetherell's discursive psychology.This study adopted both the constructionist and interpretive paradigms.The ontological and epistemological underpinnings of these paradigms provided a platform from which to engage with people's subjective experiences around what is real for them and how they can make sense of these experiences.The social constructionist approach was essential to understanding the social, historical and political lens that follows from South Africa's turbulent history.
The research sample was drawn from a large, racially representative national, privately owned organisation within the retail sector (hereafter referred to as Company X).Company X is a South African Information and Communications Technology company.The organisation, with nearly 3200 employees, has gained much recognition and is a partner to a considerable client base in both the private and public sectors.The organisation was considered appropriate for this study in terms of its racial composition as it is ranked amongst the top companies in terms of Black Economic Empowerment.Given that only one organisation was studied, a case study design was used.Yin (1984) defines a case study research method as: an empirical inquiry that investigates a contemporary phenomenon within its real-life context; when the boundaries between phenomenon and context are not clearly evident; and in which multiple sources of evidence are used.(p.23) In the context of this study, the case study was design was chosen to represent the particular organisation from which the data were collected.This is important if we are to consider Yin's guidance about context.The context of this study is deeply embedded in the culture, climate and, amongst other things, the theoretical and practical understandings of AA from the organisation's perspective.

Research procedure
As first contact, a detailed letter outlining the aims of the study was emailed to the human resources manager of Company X.The letter included the researcher's intentions, what participation would involve, the methods that would be employed in the study and issues of confidentiality and anonymity.
Once access was granted, a formal letter was received which confirmed that the researcher was allowed to conduct the study at the Company X.
Purposive sampling was used in order to ensure that there was equal representation of race, gender and job profiles.Babbie (2010) describes purposive sampling as: a type of non-probability sampling in which the units to be observed are selected on the basis of the researcher's judgment about which ones will be the most useful or representative.(p.193) A database containing the contact details and demographic details of all employees in Company X was provided to the researcher.A condensed list was drawn up which used race, organisational position and gender as criteria.This list of potential participants was then sent to the human resources manager who sent out an email to those identified on the list.The email stated that employees should expect to be contacted by a researcher and that this correspondence is acknowledged, and allowed, by Company X.After sending off the email to potential participants, each of the contacts were contacted and invited to participate in the study.If the employee was unwilling to participate, they were thanked for their time and removed from the list.If the employee agreed to participate the researcher discussed a date and time to hold the interview at the participant's convenience and provided the participant with a confirmatory email a day prior to the scheduled interview.

Participants
In total, six of the participants were black, four were white, three were mixed-race and four were Indian.The sample comprised both female (n = 12) and male (n = 5) participants.
Although the initial list consisted of equal numbers of male and female employees, when contacted, female employees were more willing to participate in the study.As a result, the sample had more women than men.Years working at Company X ranged between 1 and 18 years with an average of 7.2 years.Participants were sourced from all levels of the company ranging from the regional executive manager to administrators and call coordinators.All participants could converse in English even though English was not necessarily their first language.

Research aims
The main objective was to explore employee's social constructions of AA in a South African organisation.In doing so, this study outlined two main aims: 1.The first aim was to critically analyse the discourses around AA produced by employees within in a racially diverse, privately owned South African organisation within the retail sector. 2. The second aim was to explore how the historically advantaged and disadvantaged employees in this organisation construct AA.
In achieving these aims, the discursive resources and rhetorical arguments that employees used in talking about their perceptions and experiences of AA were investigated.

Data collection
It is important to note here that discourse analysts are not interested in the processes that take place in reality or in an individual's mind; rather, the interest is in how different versions of truth are constructed (Potter & Wetherell, 1987).Thus, when interviewing, there is the assumption that the resultant conversation arises from pre-existing resources and therefore results in different versions of 'truth'.Semistructured interviews were used to collect the data in this study.As in the case of interpretative methodologies, discourse analysts use interviews when collecting data and favour contexts that pose minimal disturbance to the natural setting (Terre Blanche & Durrheim, 1999).Discourse analysts consider the interview as an avenue from which linguistic patterns can arise, thus proving ideal for the purposes of this study.The appropriateness and value of teaming qualitative interviews with constructionist studies are noted by many authors (Phillips & Jorgensen, 2002;Terre Blanche & Durrheim, 1999;Weinberg, 2002).
Babbie (2010) describes qualitative interviewing as being based on a set of topics to be discussed in depth as opposed to standardised questions.The interview as a form of data collection allows the researcher room for active intervention (Potter & Wetherell, 1987).
There are many forms of interviews (unstructured, structured, semi-structured, etc.) within qualitative interviewing.Semistructured interviews were used in this study which allowed for more flexibility and creativity in the interview process.
Semi-structured interviews are extremely popular within the discursive psychology tradition particularly because they allow participants the opportunity to influence the direction of the interview.This allows the researcher to study the discursive patterns that are constructed by participants through the use of specific discursive resources (Phillips & Jorgensen, 2002).
Before starting the interview, a biographical questionnaire was completed with each participant.The information collected included participant gender, race, organisational position, the number of years that they have been employed by Company X and their highest level of qualification.
The development of the interview discussion questions was guided by discourse analytic theory particularly based on the guidelines presented by Potter and Wetherell (1987) for doing discursive analysis.Potter andMulkay (1985, as cited in Potter &Wetherell, 1987) suggest that the interview should be interventionist in nature and that the formal procedures that generally restrict variation in traditional interviews should be excluded.In other words, and in doing so, the researcher tried to create interpretative contexts through interview questions in a way 'that the connections between the interviewee's accounting practices and variations in functional context become clear' (Potter & Wetherell, 1987, p. 164).
The interview questions drew largely from discursive psychology as the study's theoretical framework, and the questions were therefore focused on, amongst other things, a critical engagement with historical and cultural specificity.
The questions were open-ended and tapped into areas related to AA both from a theoretical and a practical point of view.Essentially, the main aim underpinning the interview questions was to explore employees' discourses around AA and to be able to study some theoretical questions about the ways in which employees perceive, and subsequently experience, AA.Prior to each interview, the researcher sent each interviewee an email reminder which served to confirm the time and date as well as the venue for the interview.
The venue for most of the interviews (14 interviews) took place in a discussion room at Company X.The room was quiet, presented no outside interference such as noise disturbances and was very private.The door was kept closed during interviews.The other three interviews took place at different venues.The regional human resources manager and the regional executive manager were interviewed in their own offices.Both offices were quiet, and no outside interference was noted.One other participant, a field service engineer, was interviewed on the site of one of Company X's clients.The interview took place in a boardroom on the site.
No outside interference was noticed and the boardroom door was kept closed at all times.All interviews were digitally recorded.
Before commencing with the interview questions, the researcher went through an informed consent form with the participant that explained the participant's right to withdraw from the study, issues of anonymity and confidentiality and, amongst other things, the fact that the interview would be recorded.Participants were thanked at the end of each interview.Participants were also provided with the researcher's contact details should they wish to contact her for any clarification or any other research-related concern.

Content analysis
The entire research process, from the development of the research questions to the transcription and analysis of the data was conducted using discursive psychology as the interpretive tool.The analysis of the data was done primarily from guidelines for analysing discourse as outlined by Potter and Wetherell (1987) and Phillips and Jorgensen (2002).Potter and Wetherell (1987) describe the transcription of data as a constructive and conventional activity.The specificity about how detailed a transcription should be remains nebulous.For example, whilst much research has been conducted on the intonational features (such as accounting for the tones used in conversation) of discourse (Kreckel, 1981, 1982, as cited in Potter & Wetherell, 1987), Potter and Wetherell suggest that for many research questions, such detail is not essential.
All 17 interviews were transcribed to Microsoft Word ® documents.All interviews were transcribed verbatim.The interviews were transcribed using simplified notation to identify broad discursive patterns.
The goal of coding is to convert a large body of discourse into manageable chunks (Potter & Wetherell, 1987).All transcribed data was uploaded electronically to Nvivo 8 which is an electronic software package for the qualitative analysis of large bodies of textual, graphical, audio and video data.Nvivo 8 is very useful where deep levels of analysis are required.
After approximately seven readings of the transcripts through the Nvivo software, a preliminarily analysis yielded 33 themes.After the preliminary analysis, the themes were synthesised and condensed into seven 'families'.In other words, related themes were combined to form larger 'families' of ideas.The seven families were then used to construct four main discourses.This article draws on one of these discourses: constructing racial hierarchies of skill.

Findings and discussion
Before discussing the findings, there are a few important points that need to be considered.Firstly, as a central feature of discursive psychology, discourse is not to be approached as representations of mental states or actual events; rather, it depends on the broader discursive system in which it is embedded (Potter & Wetherell, 1987).In other words, when analysing the data in this study, care was taken to engage in the range of discourses that participants used, in different contexts, to justify their own accounts.
Secondly, throughout the analysis phase, 'inconsistencies' in participant's accounts were noted.For example, a participant would at times contradict an earlier statement made.Such 'inconsistencies' were used to explain how talk is context dependent because, after all, inconsistencies within discourse analysis are not seen as problematic.Another comment can be made about the ways in which participants positioned themselves, particularly in instances that seemed contradictory.For example, a participant in one instance could position himself as supportive of company policies and then in another instance as being against attempts of redress.This kind of contradiction was welcomed because discourse analysis portrays people as multidimensional since they use different resources to move between different resources, depending on the context (Potter & Wetherell, 1987).
A final comment can be made on the context in which the findings are located.In other words, the findings that follow are located with a socio-political perspective and are presented in a manner that employs 'critical reflexivity' from which to draw and with which to engage with the ideological consequences of language use (Billig, 1999;Wetherell, 1998) and the ways in which social organisation is based on unequal relations of power (Phillips & Jorgensen, 2002).
The discussion that follows thus provides a platform from which to critically engage with articulated productions of AA amongst participants in way that is appreciative of South Africa's rich socio-political history.
As noted earlier, this article draws on one particular discourse: constructing racial hierarchies of skill.
This theme is based on the finding that despite government's efforts to correct historic injustice through policies of redress, racial stereotyping remains embedded within South African society.In this case, the construction racial hierarchies of skill is explored.This is the illegitimate belief that white people are intellectually, and otherwise, superior to non-white people.
Historically under apartheid, black South Africans were classified as intellectually inferior and as 'not to be trained above certain forms of labour' (Seohatse, 2011).Black people were described as 'hewers of wood and drawers of water' (Gale, 2009) and, elsewhere, as 'too lazy and ignorant to support themselves' (Coleman, 1971, in Durrheim, 2011).These crude and obviously racist remarks functioned in constructing discourses around black incompetence and produced unequal power relations.Naturally then, black others were systematically presented as unequal to the white self.Expressing attitudes and thoughts of this kind functioned also in confirming existing power relations (Billig, 1991;Wetherell & Potter, 1992).The important point here is that such discourses become entrenched in society, in everyday practices and in life in general.It is argued that these 'embedded discourses' are presented as 'taken-forgrantedness' (Fozdar, 2008) and continue to circle South African living, in arguably more subtle and covert manners (Augustinos & Every, 2007).
These racist ways of constructing black South Africans are still embodied in everyday practices.It is argued that through everyday talk, people, whether voluntarily or not, discursively draw on resources that function in sustaining historical privilege and ideas about inferiority and superiority (Augustinos et al., 2005;Duncan, 2003;Franchi, 2003).Durrheim et al. (2011, p. 26) argue this point well by suggesting that 'it is from these regulated practices -our activities -that these forms of social life and racialised subjectivities emerge to constitute race trouble'.
Extract 1 illustrates the ways in which discursive strategies are used to reinforce stereotypes about othered groups of people.
In this particular extract, a discourse of black incompetence is at play.This commentary arose in response to a question about the merits of AA: Extract 1: W (white female): 'Yes once again you get the right man for the job, and if you train them and they still can't do it, then you've got to look at another thing and say 'okay, your limits are there ... maybe you not even interested in admin'.I can go and put you in practical, something more practical.Give something for the guy who can also enjoy it and enjoy the benefits of going further, not just put him in a place where he is totally silly and not because he is stupid, but because that's just not his, his path.I mean I know of an instance of a person that's actually put in a position -out of choice I mean, not by force -but they just cannot do the job, and of course they get nailed all the time and not getting proper increases etc.Because they can't actually, actually do the job.It's not their line, so yes it is, but if a guy cannot do it or he is not happy, move him to where he is happy and develop him from there'.
In extract 1, W's talk functions in dichotomising competency and skill on racial lines.Firstly, W's reference to getting 'the right man for the job' indirectly infers that currently, the black employee in question is not the 'right person for the job'.The speaker's emphasis on situating such employees within more 'practical' jobs serves in reinforcing stereotypes around competency along racial lines, suggesting that black employees may be more suited to the 'easier' or more 'practical' types of jobs.In many ways, the speaker is also revisiting and legitimising historic constructions of inferiority, where black employees would typically be found in positions requiring lower levels of skill.This idea contributes to the ideology of racialised competence to the extent that it 'normalises' the idea that black people are inappropriately suited to some types of work.Notice how W also credentialises (Fozdar, 2008) and softens her formulation (Edwards, 2000) about placing people in 'practical' jobs by presenting it as something good for the person concerned ('who can also enjoy it and enjoy the benefits of going further').
The speaker's choice of words 'where he is totally silly' presents 'him' as almost helpless in certain positions to the extent that he is not able to acquire the necessary skills to get the job done.The speaker's words 'not because he is stupid but because that's just not his, his path' functions in a contradictory manner to neutralise her previous comment.This suggests two things.Firstly, by asserting that he [the black employee] is not stupid, the speaker attempts to position herself as someone who does not support the idea that black employees are silly (or incompetent) by virtue of their colour.Secondly, the speaker validates and reiterates the positioning of herself as someone 'fair' by suggesting that the reason for the employee's incompetence is because it is not 'his path' [and would therefore be disadvantageous to him] rather than along lines of racial stereotyping about incompetence.It can be said here that the speaker is using anti-racist talk as a device to defend her position as someone who is concerned.This mix of racist and anti-racist talk is what Fleras (1998) refers to as a duelling discourse.Arguably, the speaker is engaging in devices that present otherwise negative views as reasonable and at the same time protects the speaker from charges of racism and prejudice (Augustinos & Every, 2007).
W's remark that 'they get nailed all the time' suggests that black incompetency is not an exception but rather is something that is frequent.This comment is presented almost as a truth and as a matter of factness, arguably as support for her position.W also comments that the black employee who 'cannot do it [the job]' should be moved to a place where he is 'happy' and 'develop him from there'.This discourse implies, although not directly, something about the trainability of black employees.Importantly, as discussed in the literature, there is an identified lack of skill amongst black employees (Chenault, 1997, in McFarlin, Coster & Mogale-Pretorius, 1999;Segwati, 1998).The lack of skill is particularly as a result of poor schooling in pre-democratic South Africa, which has inspired some of the aims of the Employment Equity Act, which promotes reasonable accommodation, training and development for people from designated groups.Despite these legal accommodations there is still reservation from people.
Moving back to W's comment above, two points require attention.Firstly, there is the inference that black employees are just not suited to some types of jobs and should therefore be moved to arguably 'simpler' jobs that would make them 'happy'; the speaker's failure to address issues of training for people 'who cannot do it' immediately disqualifies black employees from certain types of jobs.Secondly, in moving someone into a position where he is 'happy' and to a position where he can be 'developed' polarises employees along 'racially re-traced lines' (Franchi, 2003) in manner that legitimises white status and power and the exclusion and stigmatisation of black employees (Franchi, 2003).
Discursive devices used in talk, as in the case of W, function in reinforcing stereotypes about the racial hierarchy of competence.In doing so, they consequently bring into question, and undermine, the efficacy of the measures and directives of AA, which by default, through a process of inversion (Duncan, 2003), is constructed as a system that encourages black incompetence.Put simply, by undermining and discrediting AA 'candidates', AA as a system of redress is simultaneously constructed as problematic.Of importance is the fact that as a 'practice' AA is presented as problematic, but that this does not necessarily refer to the actual principles of the policy.
Of particular importance is that much of the talk in the current study often resulted in an illegitimate association being made between AA and incompetence, suggesting that they are inordinately linked in many ways.This imagined association was evident from both groups -those who were seen as benefiting from AA practices and those who were seen as 'disadvantaged' by the policy.An important point here is that the construction of skill along racial lines is not only reinforced through talk by white participants but also internalised as 'true' or 'evident' by black participants.Some supporting extracts are visited below: Extract 2: M (white female): 'How are you empowering him?You make him lazy, you make him think that you now just got the job because of the colour of your skin, whereas you actually need to study very hard as you know.It is hard studies.It is a lot of sacrifices, and that is the way how you climb the ladder, but we are really not doing those people a favour and you know they will bring their children up like that as well'.
Notice how M begins her argument with a rhetorical question which she immediately answers for herself.M's answer also functions as a duelling discourse (Augustinos & Every, 2007).On the one hand, M constructs AA as a system that makes people 'lazy' and on the other, AA is seen as a system that encourages self-doubt and personal insecurity ('you make him think that you now just got the job because of the colour of your skin').In both instances, constructing AA in this manner functions in discrediting it as a legitimate system.
There is an immediate assumption that the black employee in question is actually unqualified or incompetent because he actually got the job based on the 'colour of your [his] skin'.This kind of talk neglects the possibility that the employee in question may well in fact be skilled and competent and be black, all at the same time.Again, this reinforces an unjustified association between colour and skill which also functions to reinforce historic stereotypes about race and skill.The speaker's inferences about race and skill associations further supports her implied message that black employees are actually unsuited and unqualified for the job because you 'actually need to study very hard' (suggesting that the black employee has not in fact studied, or studied hard and, as such, has undeservingly gained from [unfair] AA practices).
M's talk in extract 2 also constructs black employees as 'lazy' (or as having the propensity to be lazy), as not working hard and, indirectly, as not able to undertake studying successfully ('It's hard studies').M's statement, 'they will bring their children up like that' reinforces racial stereotypes about parenting and constructs a particular 'uniqueness' about black people (they) and the way in which they bring up their children.This talk functions in two ways.Firstly, describing black employees using the word lazy functions as a naming tactic (Potter & Wetherell, 1987), which is a common sense device used to articulate racist discourse (Fozdor, 2008).Secondly, the (negative) outcomes of AA are construed as far reaching, expanding beyond the workplace and as a possible contributor to, or facilitator of, poor parenting ('they will bring their children up like that as well'): Extract 3 reinforces the perceived association between colour and skill and, although not stated explicitly, is enacting a kind of laissez-faire racism (Bobo, Kleugel & Smith, 1997), an ideology that supports democracy but opposes the principle of AA.Considering her emphasis on the importance of 'qualifications' and 'experience', N's talk implies that when it comes to AA practices, things such as qualifications and experience are not taken into consideration, thereby questioning the authenticity of AA as an initiative of redress and, arguably by default, implying that black employees are inexperienced and unqualified as a rule, rather than as an exception.Her talk portrays a sentiment of disappointment and, again, neglects the possibility that an employee can be qualified, experienced and black.This kind of talk once again functions in perpetuating stereotypes about race and skill by suggesting that some race groups are more likely to be highly skilled, than others.This idea is presented again in extract 4 by M, who overlooks the possibility that a person can be offered preferential treatment based on race and skill: Extract 4: M (white female): 'But yes, it is because there's some they, ja, you get employed because of your colour of your skin or because you fit the profile'.
The talk drawn from the extracts above provides a means to engage with the broader challenges related to AA, both in terms of how it is understood and how it is experienced.
The analysis shows that despite concerted attempts to enact political and social redress both from a governmental and an organisational perspective, 'many Whites see redress as an instance of a wider pattern of victimisation' (Durrheim et al., 2011) and many black people continue to feel marginalised and stigmatised.AA, addressed in the literature, continues to generate controversy (Charlton & Van Niekerk, 1994;Cohen & Sterba, 2003;DeCapua, 2010;Herdhold & Marx, 1999;Mittner, 1998).Durrheim et al. (2011) expand on this argument by stating that stereotypes about black incompetence and corruption are part of a perennial discourse regarding the inability of black people to govern in Africa.The emphasis here is thus seen as being shifted away from AA measures per se.Emphasis rather, is placed on the person, who has been constructed as the problem.
Participants, whether through experience, opinion or otherwise, continue to engage with the rhetoric of 'othering' (Riggins, 1997), particularly in relation to the ways in which their talk marginalises employees, perpetuates prejudice and 'legitimises' inequalities, specifically along racial lines.Racial stereotypes about competency and skill continue to emerge through talk and make illegitimate claims about what is true, what is constructed and what is imagined.Participants, as can be seen above, engage in multiple discursive devises and rhetorical arguments that function in rationalising a racial order of competence, which not only implicitly defends historic ideologies around white supremacy but, perhaps more importantly, also inversely functions in undermining and disqualifying AA as a system that is inherently problematic.Potter and Wetherell (1987) understand this dynamic as a resource that people use in presenting their version of the world.It is important, also, to consider that history, most notably in the form of apartheid South Africa, continues to inform the ways in which 'things occur' and how we construct the self and the other.
To say this is to agree with Burr (1995), who insists that, in adopting social constructionism, we take a critical stance towards that which we take for granted because the ways in which the world is understood are inordinately tied into history and culture.Utterances, after all, are context bound (Potter & Wetherell, 1987).'Talk' functions as a form of social practice that constructs the social world.The extracts above were therefore interpreted with this in mind.

The theoretical, methodological and practical value of the study
As its theoretical framework, the current study adopted the social constructionist approach with a particular focus on discursive psychology.The use of this approach adds to the existing body of knowledge around the usefulness of critically engaging with the embodied nature of prejudice that stems from everyday practice.There are few documented studies in South Africa that have adopted social constructionism as an approach using diverse samples.Thus, the current study has shown that it is useful to adopt a social constructionist approach within the South African context, particularly when questioning taken-for-granted knowledge and trying to understand phenomena that are historically and socially specific.The study's methodology highlights the real value of studying context in meaning-making and in studying implied inferences that underlie talk.
From a methodological perspective the current study has offered a contribution to the methods that can be used to qualitatively explore subjective, contextualised experiences of AA.The interpretative approach provided an opportunity to engage with AA as a construct in a manner that allowed for the reflection of subjective experiences.This point is especially important since subjective, contextualised approaches to AA have received little attention both locally and internationally.Practically all of AA-related research is located within the positivist paradigm.This study has also contributed the nature of the relationship between researcher and the phenomena under study, particularly as a result of the researcher's awareness of issues of reflexivity.
Another methodological contribution comes in the form of the challenges presented in discursive research, particularly when conducting research in a language other than the participants' mother tongue.Future studies should consider collecting data in the first language of participants.
Lastly, this study can potentially offer some practical value to the area of AA.Firstly, the findings may contribute to the discipline of industrial psychology, particularly in the ways in which policies around preferential treatment are conceptualised and subsequently implemented within organisations.This study offers a unique perspective on how people both understand and experience AA.This knowledge may well contribute to the ways in which organisational policy documents are conceptualised so that attempts may be made to move beyond the things that serve to perpetuate inequality within the workplace.The findings in the current study also call for attention to be paid to ensuring that organisational culture functions to create an environment that challenges the negative associations of AA in a constructive way.The findings also highlight the possibility for other organisations to consider that, despite their impressive policies of inclusion and transformation, there may exist a disjuncture between the intentions of the organisation on the one hand and the experiences of employees on the other.

Limitations and recommendations of the study
Although 17 interviews were sufficient for the purposes of this study, a larger study would have allowed for alternative, and possibly more varied, perspectives on AA.The findings highlight the importance of exploring social constructions in context.This study contextualised the findings particularly within the context of South Africa.It is thus recommend that future research consider in greater detail the culture of the organisation.In as much as discourses are embedded within socio-historical contexts, the culture of the organisation, including its values, norms and assumptions, may well contribute to the ways in which people both perceive and experience AA.
Another concern is raised in relation to what Willig (2001) refers to as epistemological reflexivity.This form of reflexivity requires the researcher to engage with questions around the research design used and, amongst other things, the content and nature of the research questions.Essentially, this type of reflexivity allows one to think about the assumptions that are made about the world and helps us to think about the implications of these assumptions on the research findings.This is an important point to consider in future studies.
The role of language in discourse is distinct; in fact, discourse is considered to be 'situated' language.Given this, had participants been interviewed in their first language, the responses might have been different.Some participants spoke English only as a second language, which could have impacted their responses or their ability to express themselves in a language other than their mother tongue.Furthermore, given that the researcher might be 'culturally' different to participants, she may have misinterpreted their accounts because of being culturally unfamiliar with their talk.Given these limitations, it is recommended that related studies consider exploring the benefits or appropriateness of conducting interviews in participants' home language.
Although the issue of language should be highlighted, the decision to conduct interviews in English is not entirely problematic.A range of studies (for example, Duncan, 2003;Franchi, 2003;Potter & Wetherell, 1987;Wetherell & Potter, 1992) in the discursive tradition have been undertaken in which interviews were in English amongst participants who did not necessarily speak English as their mother tongue.

Conclusion
This study was conducted in an attempt to explore the discourses surrounding AA amongst employees in a Durbanbased organisation.In addressing historic discrimination, AA within the employment sector is an area that is receiving much attention in South Africa.AA initiatives are heavily encouraged both at the legislative and organisational levels.
Although theoretically sound at a policy level, the practice of AA remains controversial.The ambivalent opposition to AA suggests the importance of exploring the extent to which employees embrace AA in the workplace, especially given that South Africa is a relatively new democracy with many of its inclusive policies still in their infancy.Given this, this study's main objective was to explore AA from a social constructionist perspective.Using discourse as a method of enquiry allowed for the critical engagement with the embodied nature of prejudice that stems from everyday practices.
This study had two specific aims.The first aim was to critically analyse the discourses around AA produced by employees within a racially diverse, privately owned South African organisation.The second aim was to explore the constructions of AA by historically advantaged and historically disadvantaged employees in this organisation.
The literature on AA is consistent in its emphasis on how AA continues to raise questions about equality and fairness within South Africa.
Approaching this study using social constructionism as a theoretical framework highlighted how the discursive method of inquiry proved ideal as a tool to study the pervasive, recurring patterns of talk which function to justify and rationalise historic privilege and the reproduction of social inequality.
The discourse of constructing racial hierarchies of skill was based on the finding that despite government's efforts to correct historic injustice through policies of redress, racial stereotyping still features in ways that construct racial hierarchies of skill, specifically through an illegitimate belief that white people are intellectually, and otherwise, superior to non-white people.Overall, this theme illustrated the way in which participants' talk had the potential to marginalise employees and 'legitimise' inequalities, specifically along racial lines.Also noted was the practice of shifting away from AA measures as a policy and locating the problem within the person.Again, this points to the argument that whilst participants see value in the policy, they disagree with how the policy is practised.The discussions that ensued showed also how participants engaged in multiple discursive devises that functioned in rationalising a racial order of competence which not only implicitly defended historic ideologies around white supremacy but, perhaps more importantly, also inversely functioned in undermining AA as a system that is inherently problematic.
The following are some examples of the questions posed to participants: 1. Affirmative action is a legislative requirement within South African organisations.What do you understand by the concept of affirmative action?What does affirmative action mean to you? 2. Part of the reason for implementing the affirmative action policy is to create greater opportunities for employees who, under the apartheid era, were discriminated against on the basis of race.What do you think about the 'preferential' treatment endorsed by the affirmative action policy?
Extract 3: N (white female): 'I've had to work hard to get where I am at now you know what I mean, why must I just give it all up? ... That in a company that if you are going to employ somebody you must employ them on qualifications, experience and if they deserve the job then they must get the job you know.Nowadays they don't look at it like that.They don't'.