RACISM UNDER THE LINGUISTIC MICROSCOPE: A PRELIMINARY INVESTIGATION INTO WHAT THE CRITICAL ANALYSIS OF SOUTH AFRICAN NEWS MEDIA DISCOURSE CAN REVEAL ABOUT RACIAL STEREOTYPING, AND THE RELATIONSHIP BETWEEN CULTURE AND LANGUAGE.

Abstract

A preliminary investigation into what the critical analysis of South African news media discourse can reveal about racial stereotyping, and the relationship between culture and language. Keywords: Race, Racial Stereotyping, CDA, Critical Discourse Analysis, Folk Theory, Us vs. Them, News Media, Language, Power, Linguistic Ideologies

Full text

Christine Cozien
1
RACISM UNDER THE LINGUISTIC MICROSCOPE:
A PRELIMINARY INVESTIGATION INTO WHAT
THE CRITICAL ANALYSIS OF SOUTH AFRICAN
NEWS MEDIA DISCOURSE CAN REVEAL ABOUT
RACIAL STEREOTYPING, AND THE
RELATIONSHIP BETWEEN CULTURE AND
LANGUAGE.
Christine Cozien
A research essay submitted in fulfilment of the requirements for the
degree of Honours in Linguistics, University of Cape Town
Cape Town
2015

Christine Cozien
2
ACKNOWLEDGEMENTS
I would like to thank my supervisor, Professor Rajend Mesthrie, for his guidance,
and for financial support via his NRF SARCHI chair.
I would also like to express my gratitude for financial support from the NRF, and
from UCT via the Lestrade Scholarship.

Christine Cozien
3
CONTENTS
INTRODUCTION
1.1 CRITICAL DISCOURSE ANALYSIS VERSUS DISCOURSE
ANALYSIS – WHAT’S THE DIFFERENCE?
1.1.1 CRITICAL DISCOURSE ANALYSIS DEFINED
1.1.2 CRITICISMS AND DEFENCE OF CRITICAL DISCOURSE ANALYSIS
1.1.3 METHODOLOGY OF CRITICAL DISCOURSE ANALYSIS
1.1.3.1.a WORDS (LEXICAL ANALYSIS)
1.1.3.1.b PREDICATION
1.1.3.2 SENTENCE CONSTRUCTION (1): SYNTAX AND TRANSITIVITY
1.1.3.3 SENTENCE CONSTRUCTION (2): MODALITY
1.1.3.4 PRESUPPOSITION
1.1.3.5 RHETORIC
1.1.3.5.a METAPHOR
1.1.3.5.b METONYMY
1.1.3.5.c NEOLOGISMS
1.1.3.5.d PUNS AND WORD PLAY
1.1.3.6 THE NARRATIVE
1.1.4 ADDITIONAL, RELEVANT THEORETICAL CONCEPTS
1.1.4.1 FORMALIST VS FUNCTIONALIST APPROACHES

Christine Cozien
4
1.1.4.2 SOCIAL CONSTRUCTIVISM
1.1.4.3 FOLK THEORY OF MIND, OR FOLK PSYCHOLOGY
1.1.4.4 CRITICAL LITERARY THEORY
1.1.4.5 THEORY OF COMMUNICATIVE ACTION
1.1.4.6 LINGUISTIC RELATIVITY
1.1.4.7 LINGUISTIC IDEOLOGIES
2.1 RACISM
2.1.1 DEFINITION
2.1.2 RACISM IN THE SOUTH AFRICAN CONTEXT
2.1.3 A NOTE ON LINGUISTIC DISCRIMINATION
3.1 DATA ANALYSIS
3.1.1 NEWS ARTICLE – WRITTEN DISCOURSE
3.1.2 VISUAL DISCOURSE
3.1.3 WORD FREQUENCY
3.1.4 FOLK THEORY – RACIAL STEREOTYPING AS HUMOUR
CONCLUSION
REFERENCES

Christine Cozien
5
RACISM UNDER THE LINGUISTIC MICROSCOPE: A PRELIMINARY INVESTIGATION INTO WHAT
THE CRITICAL ANALYSIS OF SOUTH AFRICAN NEWS MEDIA DISCOURSE CAN REVEAL ABOUT
RACIAL STEREOTYPING AND THE RELATIONSHIP BETWEEN CULTURE AND LANGUAGE.
INTRODUCTION
Many scholars (Sapir (1921), Whorf (1939), Duranti (2001), Halliday (1977), van Dijk
(1985,1991), Foucault (1971a), Fairclough (1989), Kress and van Leeuwen (2006)) have, over
the years, either alluded to or expressly pointed out the circular relationship between language,
culture and society. Several very different anthrolinguistic studies have all shared the premise
that “(1) language cannot be studied in isolation from social practices or cultural meaning, and
(2) questions about social relations and cultural meanings can best be answered by paying close
attention to language” (Ahearn, 2012:17). It is not only linguistic anthropologists who choose to
investigate society and culture through the linguistic lens, but sociolinguists, ethnographers and
critical discourse analysts too. Linguistics today has been emancipated from the Chomskyan/
Saussurean school of thought whereby linguistic performance/ parole is considered to be of no
interest or consequence to linguistic research. Language, culture and society are not separate
entities, but rather share a very complex multidirectional relationship. Language is our most
powerful tool for understanding how we construct and comprehend our socio-cultural world.
Racism has a long, dynamic history, and remains deeply entrenched in the thoughts and actions
of society. How people speak about race can reveal important information about the cognitions
that support racist ideologies. Furthermore, it can provide us with a deeper understanding of how
we manifest our socio-cultural reality with words, and in turn allow our language to be shaped by
that reality. Van Dijk (1993: 146) explains that racism consists of two subsystems – social and
cognitive. The social side of racism manifests on a micro level through social practices of
everyday racism, and on a macro level via power abuse by dominant groups, organisations, and
institutions. An example of macro-level racism is companies hiring staff based on race, or
politicians’ racist hate speech being reported in news media, whereas micro-level racism would
be a woman being treated as if she has no money in a store because of a racial stereotype about
wealth and her “racial group”. Discussions on social media or public forums which involve the

Christine Cozien
6
employment of racial stereotypes, such as are discussed in this paper, are a bit more complicated
to categorise in terms of van Dijk’s model, because they manifest at a micro level, but are
facilitated by a macro-level medium, i.e. the internet, and online news media. The second
subsystem of racism is cognitive. The tangible discriminatory practices of members of dominant
groups and institutions have a mental basis constructed from biased theories about ethnic events
and interactions, which in turn are based on racist prejudices and ideologies, and popularised at
micro level. Discriminatory practices are not necessarily always conscious and intentional.
Rather the problem stems from the fact that they presuppose socially shared, negatively
orientated stereotypes about ‘Us’ and ‘Them’. Racism affects all nations, cultures, and levels of
society, and South Africa is no exception. I have observed that South African daily discourse is
still largely divided along racial lines. This is evident in societal discourse on both macro and
micro level, as well as shared societal cognitions (folk theories) about race. The categories used
to racially classify, and divide, South Africans during Apartheid are still prevalent in the lexicon
of micro-level society and are actively encouraged and maintained by the macro-level of society
(viz. news media and the social actors they report on).
Inspired by work by van Dijk (1991) on racism in news media in Western Europe, this paper
aims to conduct a similar sort of preliminary investigation into South African news media. Van
Dijk’s (1984, 1987, 1991, 1993) extensive research into racism and discourse shows how news
media play a key role in keeping racism and racist discourse alive in Western Europe. This
research investigates whether these results can be generalized to South African society.
Van Dijk’s studies (1984, 1987, 1991, 1993) on racist discourse used Critical Discourse Analysis
(CDA) as their methodology. There is not one set methodology for CDA and so, for this paper, a
number of different approaches (Fairclough (1989, 1995), Stubbs (1996), Halliday (2006),
Richardson (2007)) were investigated. It became apparent that it would be necessary to consider
research by linguistic anthropologists such as Hill (2008), which introduced the idea of linguistic
ideologies and the role they may play, as well as that of Hymes (1962), Habermas (1988, see also
Fultner 2011), and Sapir (1921). The relationship between socio-cultural phenomena such as

Christine Cozien
7
racism and language is a complicated and multi-faceted one, and thus needs to be considered
from a number of different angles.
To get a good sense of the content currently being published by South African news media, and
to collect data, the following mechanisms were put in place. A wide range of print news media,
such as the Cape Times, The Argus, The Mail & Guardian, and The Times were read on a weekly
basis. Because online news media today has as much, if not more influence on society, a plethora
of online news media streams – News24, iol, Mail & Guardian, SABC News, The Sowetan and
the South African Civil Society Information Service, as well the public comment on the articles in
the feed were also observed and recorded. This live data collection was conducted on a daily
basis over a period of roughly six months. Additionally, so as not to limit data to too brief a time
period, print media articles were sourced from Sabinet dating as far back as 1994. During the
fieldwork process the following observations were made:
Firstly, the sense and psychology of racial division and difference in South African society is
perpetuated and maintained via the simple rhetorical device of repetition. A somewhat
unconventional but very effective means of visually representing this – a word cloud – has been
used in this paper. The details of how a word cloud works are explained in more detail in the
analysis but, in short, a word cloud is a visual representation of levels of word frequency in a
text, generated by a computer program.
Secondly, it appears that a folk theory exists amongst South Africans that racism and racial
stereotyping is acceptable provided that it is committed with humorous intention or is ‘tongue in
cheek’ i.e. well disguised sarcasm. So the perception of intended meaning is pivotal in
determining whether it is considered offensive or not. This ties in with Hill’s writing on slurs
versus gaffes. Two recent, popular South African publications exemplify this – The racist’s
guide to the people of South Africa by Simon Kilpatrick, and Stuff South African White People
like by Christian Lander and Hagen Engler. An extract from the latter, along with an analysis of
public comments discussing the publication, is examined in the data analysis.

Christine Cozien
8
The methods of investigation mentioned thus far are the author’s own. As mentioned in the title,
this investigation is preliminary and investigative in nature, and it is hoped that these methods of
data analysis may contribute to the field of Critical Discourse Analysis.
Beyond this, methods for Critical Discourse Analysis laid out in Richardson’s Analysing
Newspapers (2007), together with some of those outlined in Fairclough’s Language and Power
(on which parts of Richardson’s work is based), were used to analyse one full newspaper article.
According to Nursya’bani et al (2013) “Critical Discourse Analysis explores the connection
between the use of language and the social and political contexts in which it occurs”. Because
the goal of this research is to gain some understanding of the relationship between language use
in South African news media and racism in South African societal discourse, Critical Discourse
Analysis has been chosen as the methodology for the investigation.
Discourse is not only transmitted orally or aurally, but also visually, and for this reason an
image, published in several forms of news media in South African, has been included, and I have
analysed this using methods laid out by Kress and van Leeuwen, in their work Reading Images
(1990).
Restrictions on the length of this paper place limits on the amount of data that can possibly be
analysed and so, with the goal of perusing some of the many possible methodologies available
for the analysis of news media discourse, each piece of data was analysed using a different
methodological approach.
This work has an agenda of social analysis, and does not limit itself to a purely grammatical
analysis of discourse. It thus falls under the banner of Critical Discourse Analysis, and not
Discourse Analysis.

Christine Cozien
9
1.1 CRITICAL DISCOURSE ANALYSIS VERSUS DISCOURSE ANALYSIS – WHAT’S THE
DIFFERENCE?
There remains, within the field, little uniformity on the use of the terms discourse analysis and
critical discourse analysis, and little clarity on the meaning and the distinction between the two.
Discourse analysis should refer to the analysis of discourse, performed with the aim of gaining
understanding about the grammar and (thus) structure of language. Formalist (bottom up) and
functionalist (top down) approaches and methodologies are both appropriate for discourse
analysis. However, once we conduct discourse analysis with the aim of understanding speaker
motivation, and wish to hypothesise about social and cognitive motivations for discourse, I
believe that this implies an element of critical social theory, and this then is Critical Discourse
Analysis (CDA). This viewpoint follows from Janks’ (1997:329) statement that “where analysis
seeks to understand how discourse is implicated in relations of power it is called critical
discourse analysis.” The Formalist approach is not at all suited to CDA, whereas Functionalist
approaches and methodologies offer a valuable point of departure. A more detailed discussion of
Formalism and Functionalism appears later in this paper, under 1.1.4.1 Formalist versus
Functionalist Approaches.
1.1.1 Critical Discourse Analysis defined
The following two quotations neatly summarise the approach of Critical Discourse Analysis:
“Critical Discourse Analysis explores the connection between the use of language and the social
and political contexts in which it occurs” (Nursya’bani et al, 2013)
“CDA argues that there are relations between language, power and ideology, and between how
the world is represented in texts and how people think about the world” (Stubbs, 1997:2)
“The overall aim of CDA has been to ‘link linguistic analysis to social analysis’” (Richardson
2007:26 citing Woods & Kroger 2000:206)

Christine Cozien
10
CDA remains a developing field and there are a number of different offshoots; groups of
scholars who have proposed different ideological and methodological approaches. Breeze
(2011:494) breaks it down neatly, distinguishing “between the initial British approaches
embodied by Fairclough (1985, 1989) and Fowler (1991); more developed and coherent form
explained in Chouliaraki and Fairclough (1999); the so-called “sociocognitive model” of critical
discourse analysis epitomised by van Dijk (1991) and his group; and the Viennese “discourse
historical school” led by Wodak (Wodak et al. 1990; Wodak 1996, 2007)”. This list is by no
means exhaustive and mention must also be given to scholars such as Hymes, for his
‘Ethnography of Speaking’ model (1962), and Halliday for his work on Systemic Functional
Grammar (1977).
Nursya’bani et al (2013) summarise four key tenets of CDA. Firstly, that social and political
issues are constructed and reflected in discourse. Secondly, that power relations are negotiated
and performed through discourse. Thirdly, that discourse both reflects and reproduces social
relations, and lastly, that ideologies are both produced and reflected through discourse.
Richardson (2007:46) mentions three key issues in CDA which must be considered in the
analysis of discourse. The first is that texts are written or composed in different ways to mean
different things (i.e. content and form are interrelated). In addition to that Richardson describes
discourse as having levels of analysis. This is a reference to Fairclough’s idea that there are two
main features of texts to consider during analysis: “the first has to do with the structuring of
propositions, the second with the combination and sequencing of propositions”. (Fairclough
1995:104 quoted by Richardson 2007:46). The structuring of propositions affects the way in
which individuals and social actors are represented and, in the same way, the construction of
clauses influences the representation of events, actions and processes. Essentially the levels of
analysis show how clauses are organised to create a coherent text. The third point that
Richardson emphasises is that texts are multi-functional. Here Richardson is referring to
Halliday’s (1977) Systemic Functional Linguistic theory, which assumes that texts fulfill an
ideational function, an interpersonal function, and a textual function. Texts may simultaneously
represent aspects of the physical, social and mental world – this is an ideational function. They

Christine Cozien
11
may also enact social relations between participants as well as their attitudes, desires and values,
which is an inter-personal function. The textual function of texts simply refers to the manner in
which they are connected as a cohesive and coherent whole.
1.1.2 Criticisms and defense of Critical Discourse Analysis
Critical Discourse Analysis has been criticised for being overly subjective, lacking in systematic,
empirical methodology (van Noppen, 2004, Widdowson, 1998), and for focusing on “negative”
topics (Martin 2004; Luke 2002).Widdowson (1998), in his critique of what he deems to be “an
authoritative representation...of CDA”, constantly alludes to an accusation that CDA is just
glorified literary criticism. Critical Discourse Analysis would be literary criticism if it only
critiqued literature, but it doesn’t, because discourse is constituted of more than just text: it is
visual, oral and aural too. CDA critiques all these aspects of discourse as well as the socio-
political circumstances that create, or are created by, discourse. CDA has strong links with
Critical Literary Theory and Critical Social Theory, a subject on which is elaborated upon under
1.1.4.3 Folk theory and Critical Literary Theory.
Van Noppen (2004:108) questions “whether (CDA) contributes new insights and whether it
actually helps ‘change the world’”. How he proposes we should know the answer to this is not
clear, but I think it is fair to say that all research, by its very nature, contributes new insights.
CDA is a comparatively young discipline which still needs more time to review and refine its
methodologies. Many publications on CDA are, as critics such as van Noppen (2004) and
Widdowson (1998) point out, very vague on actual concrete, applied methods for analysis. I too,
through this research, have come up against the problem posed by a lack of tangible, definite
instructions regarding methods for CDA. It is for this reason Richardson’s (2007) methods for
analysing newspapers are described in detail, and methods for analysing images and using word
frequency in analysis are put forward.

Christine Cozien
12
Breeze (2011: 493) reports that Martin (2004) and Luke (2002) have both criticised CDA
because of “the largely negative nature of the body of work produced within the field”, and call
for “critical scholars to pay more attention to positive or potentially transformative uses of
discourse”. Critical Discourse Analysts seek to expose potentially negative aspects of society
with the positive goal of creating awareness and instigating discussion on why, for example
racism or sexism, is wrong. Highlighting instances of discrimination, harassment and
marginalisation in society forces people to enter into debate, to critically analyse society, to
dispute systems or hierarchies which perpetuate the problem, and to challenge those in power to
effect neccessary social change. Creating awareness is the first step towards social
transformation, and CDA is part of that initial step.
1.1.3 Methodology of Critical Discourse Analysis
What follows is a practical methodology for the analysis of discourse, specifically text from
news media, based largely on work by John E. Richardson (2007), which is coherent technical
framework for the analysis of text discourse. This has been augmented with ideas from
Fairclough’s Language and Power (1985).
Richardson (2007:47) proposes that the Linguistic analysis of news-text should move from the
micro-analysis of words, through sentences and onto macro-analysis of the organisation of
meaning across the text as a whole, as shown is figure 1.1 on the following page.

Christine Cozien
13
LEVELS OF ANALYSIS
Structuring of Propositions Words (Lexis) Micro-textual
Sentences 1 (Syntax and transitivity) analysis
Sentences 2 (Modality)
Presupposition
Combining Propositions Rhetoric Macro-textual
Narrative analysis
Figure 1.1 Richardson’s Levels of Analysis (2007:47)
1.1.3.1.a (Words) Lexical Analysis
This is the first stage of analysis. Words convey value judgments of society through connoted
and denoted meaning. All content words, especially verbs, adverbs, nouns and adjectives, carry
connoted (suggested or implied meaning) and denoted meaning (the literal meaning or concept
we have decided that word represents). Nouns, or naming words, are often used to build an ‘us
vs. them’. There are a great number of words that may be used to describe an event or an
individual, which are all accurate but which do not all carry the same meaning (connoted or
denoted; also referential meaning and indexical or social meaning). By using one social category
to describe a character we either exclude them for other categories or choose to foreground that
social category to one or another end. Reisgel and Wodak (2001 cited in Richardson 2007:49)
call these naming options a text’s ‘referential strategies’, explaining that opting to describe or
categorise an individual or group in a particular way ‘can serve many different psychological,
social or political purposes’. A social actor may be individualised in order to emphasise his or
her ordinary, “everyday-man” qualities e.g. ‘Nelson Mandela confided’ or ‘Jacob Zuma
expressed’. A social actor may also be collectivised or categorised, thereby being attributed
different implicit and explicit meanings e.g. Comrade Nelson Mandela, Freedom Fighter Nelson
Mandela, father, communist, etc.

Christine Cozien
14
A squared relationship will usually exist between characters in a victim-vs-villain, or us-vs-them
narrative. This concept is known Teun van Dijk’s ideological square (see figure 1.2 below)
which involves positive self-presentation and negative other-presentation. According to the
ideological square ‘they’ (i.e. the out group) will be represented in a negative way, by
emphasising or foregrounding ‘their’ negative attributes and social activities, and by de-
emphasing or backgrounding ‘their’ positive attributes and social activities. Conversely ‘we’ or
‘us’ (i.e. in group members) will be represented in a positive way, using the reverse of the
aforementioned strategy.
Bad men (sex fiend) attack innocent women (mum, daughter)
Bad women (fortune hunter) provoke innocent men (hubby, family man)
Figure 1.2 van Dijk’s ideological square
The ideological square is observable across all linguistic dimensions of a text and is strongly
linked with the development of the ‘us vs. them’ narrative.
1.1.3.1.b Predication
Predication refers to assigning attributes to a noun or noun phrase with the purpose of
influencing the value judgments made about that social actor or event. Reisigl and Wodak
(2001:54) explain that “predicational strategies are mainly realised by specific forms of reference
(based on explicit denotation as well as more or less implicit connotation) by attributes (in the
form of adjectives, appositions, prepositional phrases, relative clauses, conjunctional clauses,
infinitive clauses and participial clauses or groups), by predicates or predicative nouns/
adjectives/ pronouns, by collocations or explicit comparisons, similes, metaphors, and other
rhetorical figures... and by more or less implicit allusions, evocations and presuppositions/
implications”. Predicational strategies may bestow qualities or desirability or undesirability, and

Christine Cozien
15
may be used to criticise or undermine in the same way as they may be used to support or
compliment.
1.1.3.2 Sentence construction (1): syntax and transitivity
Transitivity, in David Crystal’s Dictionary of Linguistics and Phonetics (2008), is defined as a
“category used in the grammatical analysis of clause or sentence constructions, with particular
reference to the verb’s relationship to dependent elements of structure.” A verb which can take a
direct object (as in He saw the dog) is transitive, and one which cannot (as in *He arrived a ball)
is intransitive. Some verbs can take on a transitive quality under some circumstances, or
intransitive under different circumstances.
Simpson (1993: 88) illustrates that, in any reporting, there are two main aspects of discourse that
can be changed in order to manipulate the message that is put across. One way involves the
participants involved, usually conveyed by nouns or noun phrases. We have discussed this in
section 1.1.1 (Words) Lexical Analysis. The second way in which the message can be
manipulated is via the process itself, so through the verb or verb phrase. In English there are four
principle kinds of process that can be described by a verb in a sentence. Firstly active processes,
such as speaking, singing or shouting. Secondly mental processes, such as thinking, dreaming or
deciding. Thirdly relational processes such have, seem, be; and verbs which involve an agent
and an attribute, e.g. You are x, I have y. And fourthly material processes, which are further
subdivided into transitive and intransitive processes. Transitive processes involve two or more
participants, the agent and the object of the action. (e.g. I kicked you) Intransitive processes
involve only one participant (e.g. Mc Carthy scored)
Active agent deletion often occurs during the transformation of a verb from active to passive.
Usually the passive verb without an agent is favoured in news headlines, combined with article
deletion, because it lends a certain amount of ambiguity and tempts the reader to read on in
search of more detail. An example of active agent deletion appears in figure 1.3 on the following
page.

Christine Cozien
16
 Active John kicked the ball Passive: The ball was kicked by John
 Active John kicked the ball Passive: The ball was kicked (Agent deleted)
Figure 1.3 Active agent deletion
Active agent deletion is not limited to the transitive action processes, it can be applied to active
and mental processes too e.g. ‘allegations were made’ or ‘the proposal is being considered’.
“Any transformation of this kind removes a sense of specificity and precision from the clause”
(Richardson 2007:55) and denies responsibility for the action/s.
Sometimes agent deletion occurs merely because the agent has been mentioned in the preceding
sentence, and not because of evasive intentions. This is one example of why text must be
analysed as whole, in terms of cohesion, rather than as isolated sentences. As Richardson
2007:58 emphasises, “with discourse analysis, the unit of analysis is the text as a whole and
hence analysis should ideally combine a discussion of what the text represents (includes) and
excludes – or, how the textual meaning is communicated through relations of presence and
absence – rather than on a sentence-by-sentence basis.”
1.1.3.3 Sentence construction (2): Modality
Modality is a counterpart of transitivity. It refers to “a speaker’s attitude towards, or opinion
about, the truth of a proposition expressed...or towards the situation or event described”
(Simpson 1993:47) Modality can be expressed through modal verbs (can, could, should, would,
might, may, will, must), their negations (can’t, couldn’t, etc.) or by adverbs (certainly, possibly,
etc.). Need and dare can also behave as modal verbs in English, which can be inverted into
interrogatives and can take negative markers. Modality is usually a marker of opinionated genres
of journalism, which do not merely report information, but also express value judgments of an
event or character.

Christine Cozien
17
Modality can be expressed in two main forms. Firstly through truth modality, which refers to the
degree to which the speaker believes something will or can happen. An author may choose from
a range of options, from absolutely categorical (e.g. the bloodshed will be stopped), through
varying degrees of hedging (e.g. the bloodshed could be stopped) and reduced certainty (e.g. the
bloodshed could be stopped if only...) The other form in which modality is expressed is known as
obligation modality. This refers to future events and the degree to which the speaker/ writer feels
that something should or ought to be done. Here, again, the range of options varies from
categorical (e.g. The minister of foreign affairs must demand an explanation from the Israeli
ambassador) to more cautious (e.g. .The minister of foreign affairs ought to demand an
explanation from the Israeli ambassador)
Categorical modal claims tend to be used more frequently, particularly in more sensationalized
or vigorously argued copy because they come across as more authoritative than hedged claims.
Often a hedged modal claim read by one person will turn into a categorical modal claim when
they relay the story to another person, because people like to sound well informed or want to
avoid any weak points in their story for the sake of their own integrity. A form of broken
telephone, this results in a completely unsubstantiated claim becoming popular “fact”.
Lower degrees of modal commitment can also be effective in that the ambiguity of the statement
can induce a feeling of imminent threat/ disaster/ danger or uncertainty which leads to concern or
mass hysteria and speculation (and a hike in newspaper sales as people rush to arm themselves
with more information on the topic).
Modality allows journalists, politicians and other macro-society members to make claims that are
so couched, hedged and mitigated as to be devoid of any real content or meaning at all. In order
to come across as more authoritative, journalists may sometimes chose to walk a fine line, and
replace a modal verb with its raw form, even if this means that the statement leaves a lot open to
interpretation. An example of this is the headline Now it’s water shedding (Sanpath 2014) which
ran in The Sunday Independent. The immediate assumption for South Africans reading this

Christine Cozien
18
headline, who were at that time experiencing what they call load shedding (supplier imposed
power blackouts), is that cuts to their water supply were a certainty. However, upon reading the
article one discovers that the water supplier had only warned that there might be restrictions
imposed on water supply if dam levels fell to a certain amount. Because load shedding affected
the entire country, and the water restrictions are likened to this through predication (water
shedding) the headline also leads one to believe that water restrictions would affect the entire
nation. However, again, upon reading the article one discovers that in fact, if the restrictions
come into effect, they would probably only affect parts of Durban. What sounded like a national
crisis in the headline, proves in the article to be an isolated situation which is a possibility, but by
no means a certainty.
1.1.3.4 Presupposition
Not all meaning is manifest and immediately obvious in the reading of a text. Hidden and
presupposed meaning is often present too. Richardson (2007:63) describes presupposition as “a
taken-for-granted, implicit claim embedded within the explicit meaning of a text or an
utterance”. Presupposition can take a number of different forms in texts. Reah (2002:106) lists
three linguistic structures which commonly host presupposed meaning. These are change of state
verbs (e.g. stop, begin, continue) or implicative verbs (e.g. manage, forget), the definite article
and possessive articles, and Wh-questions. For example the use of the definite article in the
phrase, the challenge facing the modern world, implies a presupposition that a challenge exists
and also that a modern world exists. Another example of is seen in the headline Who is
responsible for the poor state of the economy? in which a Wh-question implies the
presupposition that both that the economy is in a poor state, and that that some individual is
responsible for the state of said economy.
Synthetic personalisation is another way in which presupposition manifests in mass media. This
involves using inclusive language, most commonly second person pronouns, both plural and
singular, to give the masses the feeling that they are being treated as individuals. The beauty of
the pronoun ‘you’ in English is that it may refer to one or many people, in the same position in a

Christine Cozien
19
sentence. Because of this big corporations are able to both personalise and generalise their
slogans at the same time. Many large corporation today employ the linguistic strategy of
synthetic personalisation in the hope of developing ‘personal’ relationships with their clients/
customers/ consumers, and fostering brand loyalty. Common examples How can we help you?
(FNB) Inspired by you (Pick n Pay) Today, Tomorrow, Together (ABSA).
For an utterance to be understood, the presupposition needs to be mutually assumed by both
sender and receiver. This is always the case, regardless of whether the utterance takes the form of
an assertion, a denial or a question. Negation of the utterance has no effect on presupposition.
e.g. I want to see that again vs I don’t want to see that again both presuppose that the subject has
seen whatever that is before, one or several times.
1.1.3.5 Rhetoric
1.1.3.5.a Metaphor
Metaphor is defined by the Chambers 20th Century Dictionary as “a figure of speech by which a
thing is spoken of as being that which it only resembles, as when a ferocious man is called a
“tiger”. Richardson (2007:66) explains that metaphor involves perceiving one thing in terms of
another. Certain types of metaphor are associated with specific genres of journalism. For
example, metaphors of war are often used alongside sports reporting.
1.1.3.5.b Metonymy
A form of substitution in which something associated with X is substituted for X. Referring
again to the Chambers 20th Century Dictionary, metonymy is “a trope in which the name of one
thing is put for that of another related to it.” Metonymy differs from metaphor in that metaphors
operate through transference of similar characteristics, whereas metonymy operates through
replacements.

Christine Cozien
20
For example, the cause or creator may be replaced by the product e.g. The B-BBEE
1
ammendment act marginalises whites. Alternatively the user of an object may be replaced by the
object itself e.g. Girl killed by police car! Other popular uses of metonymy are where people are
replaced by the place where they work or live e.g. Parliament kicks out Julius Malema, events
are replaced by the date on which the event occurred e.g. 9/11 or June 16, or a country or state is
replaced by certain people living in that country e.g. South Africa needs to create jobs.
It is not the presence of metonyms in text that is of significance to CDA, but rather how they are
used. Metonyms may be used when the social actors responsible are unknown, or to provide a
semantic screen for responsible, affected or involved actors (victims or perpetrators) and keep
them in the semantic background
1.1.3.5.c Neologisms
A neologism is a recently coined word, or an existing word that has acquired a new, different
meaning. Neologisms are created in different ways. According to Richardson (2007:69) there
are 3 kinds of neologism that are of interest in to the study of news media. Firstly, neologisms
created by the addition of prefixes or suffixes to words e.g –gate to indicate scandal and/or
corruption, as in Nkandlagate. Secondly, neologisms created through semantic shift, either
through change in semantic function (e.g. Google noun vs. google verb), or through
euphemisation (e.g. engage ‘attack’ or neutralise ‘kill’). And thirdly, neologisms created by
combining two existing words (e.g. ‘breakfast’ + ‘lunch’ = brunch) which is also known as
portmanteau.
1
Broad-Based Black Economic Empowerment, or B-BBEE, is an economic empowerment act initiated by the South
African government. The goal of Broad-Based Empowerment is to distribute wealth across as broad a spectrum of
previously disadvantaged South African society as possible. B-BBEE is governed by the The Broad-Based Black
Economic Empowerment Act (53/2003), and employers must comply through employing according quotas relating
to race, gender and disability, if they want to be B-BBEE certified. For more information refer to
https://www.thedti.gov.za/economic_empowerment/bee.jsp

Christine Cozien
21
1.1.3.5.d Puns and word play
There are three main categories of pun. Homographic puns take advantage of words which are
spelt the same but which have different meanings and pronunciations. As a result this kind of
pun would be more common in written text or images than in speech. Ideographic puns substitute
words of similar but not identical sound. And homophonic puns substitute words with the same
sound but unrelated meanings.
Most obviously, puns are used to amuse or entertain. In news media this is often done at the
expense of some social actor, who is ridiculed by the pun. Like all rhetorical devices, puns give
away something about the publications editorial and political agendas.
1.1.3.6 The narrative
Generally, the narrative structure of any work, regardless of the medium, contains a plot, theme,
and resolution, or exposition-development-climax-denouement – these story arc elements are
simply arranged in whatever order best suits the author’s purposes. In news media the order in
which events actually occurred is not always the order in which those events are presented to the
reader.
The traditional narrative structure, as described by Aristotle, develops as follows: introduction of
characters and setting, rising action, introduction of complication, climax in which complication
is overcome, final resolution. However, because news media usually reports stories as they
develop in real-time, the narrative of news media often engages a simpler narrative, which will
be missing some of the traditional elements. Notably, the climax and the final resolution may be
missing, or the story may altered in such a way that a substitute climax and outcome are created
for the time being, but their inconclusive nature will be highlighted, leaving the reader anxious
for further, more conclusive information. News narrative may also be structured according to the
inverted pyramid structure, whereby the climax of the story is presented first.

Christine Cozien
22
1.1.4 ADDITIONAL, RELEVANT THEORETICAL CONCEPTS
An investigation of language and socio-cultural reality necessitates the consideration of concepts
such as Critical Theory (both literary and social), Social Constructivism, Formalist versus
Functionalist approaches, Folk Theory, and Linguistic Ideologies, as well as theories such as
Whorf’s theory of Linguistic relativity and Habermas’ Theory of Communicative Action. The
aforementioned are thus discussed in relative detail in the paragraphs that follow.
1.1.4.1 Formalist vs Functionalist approaches
A number of scholars in the field (Schiffrin 1994, Stubbs 1996, van Dijk 1985, Richardson 2007)
have highlighted the importance of considering the differences between formalist and
functionalist approaches to linguistic data, when undertaking any kind of discourse analysis.
As Schiffrin (1994:20) points out, the debate surrounding the nature of discourse analysis, and
the two different definitions of what it is, arise from the existence of two main “paradigms in
linguistics...(which) make different asumptions about the goals of linguistic theory, the methods
for studying language, and the nature of data empirical evidence.” These two main paradigms, or
schools of thought, have been known by a number of different names over the years. Schiffrin
explains: “What Newmeyer (1983) calls a formalist paradigm is similar to Hymes’ (1947b)
structuralist paradigm and to what Hopper (1998) calls a priori grammar; the functionalist
paradigm is sometimes also called emergent (Hopper) or interactive (Mey et al. 1992)”. There
are slight differences between all the aformentioned theories, but for the sake of brevity they are
most often grouped under the umbrellas of Formalist and Functionalist approaches.
The Formalist approach defines discourse as language above the sentence. Formalist analysis of
discourse views language as being made up of smaller linguistic units which have a particular,
hierachical relationship with one another, and which is rule based. While formalist and
structuralist approaches are often thought to be synonymous, there is a very subtle difference.
Structuralism describes grammar in terms of grammatical structures or categories, such as
morphemes, clauses and sentences. The Formalist approach expands on this by presenting a
formal model of grammar, which is described in terms of these gramatical categories.

Christine Cozien
23
In Formalist approaches, the analysis of language code or structure precedes the analysis of use.
Attempting to fit natural linguistic data into a predetermined grammatical framework can prove
to be troublesome at times, and for this reason formalists often create sentences that prove a
grammatical point.
Functionalists, on the other hand, analyse language use in natural linguistic corpora with the aim
of developing a linguistic structure or framework based on their findings. The functionalist
approach takes the structural categories of language into account, but aims to understand the
structure of language in the context of discourse and everyday language use. “Functionalists
maintain that the communicative situation motivates, constrains, explains, or otherwise
determines grammatical structure” (Nichols 1984:97).
Formalists construct a model of grammar and then attempt to fit language into that pre-
determined framework, whereas functionalists chose rather to observe language in use and then
develop a grammatical model based on that live linguistic data. In the formalist approach the
grammatical model is the point of departure, whereas in the functionalist approach it is the
destination, what they aim to discover.
Schriffin (1994:21-22) points out two further distinctions. Firstly, “formalists (e.g. Chomsky)
tend to regard language as a mental phenomenon. Functionalists (e.g. Halliday) tend to regard it
primarily as a societal phenomenon.” Secondly, “Formalists study language as an autonomous
system, whereas functionalists study it in relation to its social function”.
Formalist and functionalist approaches have conflicting ideas regarding what is acceptable in
terms of data and data analysis, or methodology. According to Stubbs (1996:23-24) formalists
believe that linguistic analysis “ can be based on intuitive data and isolated sentences (and) that
corpus data are unrevealing”. Conversely, functionalists such as Stubbs (1996:28) assert that
“language should be studied in actual, attested, authentic instances of use, not as intuitive,

Christine Cozien
24
invented, isolated sentences”. In saying this Stubbs aligns himself with a host of other
functionalists, including Firth, Halliday and Sinclair.
Functionalists find the formalist approach not only limited in that it utilises an artifical and
restricted database, but also inadequate because it is based on a idealised model of language
which is most often not applicable to natural linguistic data or discourse – ordinary people
usually do not speak in grammatically perfect or complete sentences, and aspects of language
such as semantics and pragmatics are highly context dependant.
1.1.4.2 Social Constructivism
Social Constructivism is sociological theory, most often attributed to social psychologist Lev
Vygotsky (1978), whereby groups collaboratively construct knowledge and cognitions of shared
meaning, such as folk theories. Social constructivism is the social framework by which people
learn how to be part of a particular social group or culture. The influence of culture on the
cognitive development of individuals is strongly emphasised.
1.1.4.3 Folk Theory of Mind, or Folk Psychology
Social Psychologist Betram Malle describes the folk theory of mind as “a sophisticated
conceptual framework that relates different mental states to each other and connects them to
behavior” (Malle 2006:1). In other words folk theory provides a cognitive framework which we
use to explain and understand the behaviour of others.
Social psychologists Susan Gelman and Cristina Legare, express a similar sentiment , in their
work on folk theory: “Intuitive (folk) “theories” that people construct, including theories of
mind, of biology, or of physics...are not scientific theories–they are not formal, explicit, precise,
or experimentally tested...but like scientific theories, intuitive (folk) theories have broad
implications: they organize experience, generate inferences, guide learning, and influence
behavior and social interactions. Most centrally, intuitive theories are causal and...explanatory

Christine Cozien
25
systems of knowledge...integral to human cognition and learning... Intuitive (folk) theories are
not neutral or passive snapshots of experience; they embody cognitive biases that influence
thought and action.” (2011: PubMed Central online).
In sum Folk theory is a label for collective, everyday understandings that people have about the
world. All societies have their own folk theories by which they describe and understand various
phenomena of their reality. The folk theories of one society or culture may be shared with
another, or completely opposed.
Malle (2006:1) purports that “a full understanding of social cognition must consider the folk
theory of mind as the conceptual underpinning of all conscious and unconscious perception and
thinking about the social world”. The study of Folk Theory of mind is integral to our
understanding of racism, because it is via these mental frameworks that people justify or reject
racist talk and behaviour, or judge certain kinds of racism as being acceptable while other kinds
are not.
Research on the Folk Theory of Mind, or Folk Psychology, has long been dominated by social
psychologists, but anthropologists, linguists and other scholars have also contributed to our
understanding of folk theories through their work, commonly through ethnographic analysis. The
original goal of ethno-poetic research is to understand the semantic classification systems of
linguistic communities, but, as a result, it also reveals these communities’ socio-cultural
cognitive frameworks, which they use to classify and understand the world around them.
Renowned linguists Dell Hymes and Edward Sapir are two scholars, among many, who have
studied the traditional folk stories, song and prose of speech communities to gain a deeper
understanding of their perception of reality, and how that in turn affects their linguistic
classification systems.

Christine Cozien
26
1.1.4.4 Critical Literary Theory
The term Critical Theory really has two different meanings, and it is more correct to speak of
Critical Social Theory and Critical Literary Theory. The two overlap in that they both promote
the reflective critique of society via the application of knowledge gained from research done in
the social sciences and the humanities. Critical Discourse analysis is a prime example of this.
In the context of literary studies, critical theory is interpreted knowledge which aims to decode
the meaning of human linguistic interaction. In a different sense, critical theory is a socialist,
Marxist ideology which aims to enlighten and emancipate humanity from systems of dominance
and dependence, be they social, cognitive or economic, via knowledge gained from self-
reflection.
Essentially, critical social theory seeks to challenge and critique the ideologies and discourses
within society, whereas critical theory in literary studies seeks to understand how these
ideologies and discourses affect, and play out in, text and talk. Because research in the field of
Critical Discourse Analysis analyses the language and discourses surrounding a social
phenomenon (such as racism, feminism, or sexism), elements of critical social theory become
unavoidable in such work, and so critical social theory and critical literary theory often co-exist
in critical discourse analytic work.
1.1.4.5 Theory of Communicative Action
Jurgen Habermas, a sociologist and philosopher in the school of Critical Theory, developed his
Theory of Communicative Action with the aim of grounding the social sciences in language
(Habermas 1988). According to the Theory of Communicative Action “human action and
understanding can be fruitfully analysed as having a linguistic structure...and these linguistic
structures of communication can be used to establish a normative understanding of society”
(Fultner 2011:4, 55-56). Habermas’ theory conceptualized language as a tool for passing on
cultural knowledge and generating shared meaning (folk theories), which then facilitates
solidarity. He believed that people construct their identities through communicative action.

Christine Cozien
27
Habermas believed that his speech impediment lead him to think differently about
communication. (Habermas 2008).
In terms of this research, Habermas’ theory of communicative action is enlightening in that is
suggests how people construct racial stereotypes through communicative action. Language is the
medium by which people share their understanding of different racial stereotypes and, in turn,
mould or shape shared cognitions about the nature of these stereotypes.
1.1.4.6 Linguistic Relativity
Although it is often referred to as Sapir-Whorfian theory, Linguistic Relativity is more correctly
Whorfian theory, because it was Whorf who wrote on the topic in the 1930’s and not his mentor,
Sapir (Carrol, 1956). Whorf’s theory of Linguistic Relativity contends that the linguistic
categories and patterns of a language influence the cognitions of its speakers and their
conception of reality. Whorf studied Native American Languages in an attempt to decipher how
differences in grammatical structures and language use influenced their speaker’s perception of
their world. Sapir, in fact held quite a contrary view, stating that “It is easy to show that language
and culture are not intrinsically associated” (1921:213-214).
Parallels are often drawn between Whorf’s work and Vygotsky’s.
1.1.4.7 Linguistic ideologies
Hill (2008) uses the term discourse “as shorthand for all varieties of text and talk” and not in the
ideological sense, in which discourse is line of thought or understanding about the material
world. However, she later concedes that she does “not mean just the material surface of
language” but includes “the invitations and clues, the silences, the inferences”. Inferences and
clues involve thought based on an ideological frame of reference, so the lines between ideology
and discourse become rather blurred. Foucault and others have used ‘discourse’ in the sense that
it refers to all types of discussion about a particular topic, be it thought, conversation, text,

Christine Cozien
28
images, film. The terms ‘ideology’ and ‘discourse’ seem to have become interchangeable for
many scholars, essentially because ideology i.e. the line of thought about a particular subject e.g.
communism, is inseparable from all kinds of discourse about it e.g. thoughts, beliefs, talk, text,
action and visual representation. In this sense, ideology and discourse become synonymous.
Ideologies are models by which people understand the world around them, we build, describe,
spread, challenge and destroy ideologies with discourse. Ideologies and discourse are mutually
inclusive - how else do people become aware of ideologies except through discourse?
Language ideologies are thought frameworks that rationalize certain forms of text and talk.
According to Silverstein (1979:193) “linguistic ideologies rationalize and justify what people
understand to be the structures and the prescribed use of their language/s”. Similarly Hill
(2008:32) describes linguistic ideologies as “sets of interested positions about language that
present themselves as forms of common sense that rationalize and justify the forms and functions
of text and talk.”
People acquire linguistic ideologies because they make their world more comprehensible and
coherent and because they promote and assert their access to important resources, such as
employment and education. Rumsey (1990) points out that linguistic ideologies are often
presented as common sense. Common sense obtains its status of correctness from being so
defined by a group of people whose interests are advanced by believing in it, not because it is
necessarily true.
Some descriptions and explanations of different linguistic ideologies follow:
Standard language ideology purports that there are correct and incorrect forms and uses of
language. This may also be referred to as prescriptivism. In terms of this ideology prestige
equates with correctness. A standard language will, therefore, usually be “endowed with a
respectable history and a set of arguments that explain why it is prestigious and correct” (Hill
2008:35). What is prescribed as standard language changes over time, and what may be

Christine Cozien
29
prescribed as correct or prestigious language at one time may seem less credible several
generations later. A simple example of this is certain ways of speaking about women which
were, at one time in our world’s history, quite acceptable but which now would be considered
sexist. Because standard linguistic ideologies determine what is correct and incorrect, it is in
terms of this ideology that people understand different racist stereotypes as correct or incorrect.
Personalist language ideology may be described as Humpty Dumpty logic, which is a reference
to Lewis Carroll’s Alice in Wonderland character’s statement that “When I use a word, it means
just what I choose it to mean – neither more nor less”. Personalist language ideology (Rosaldo
1981 cited in Hill 2008:39) holds that the most important aspects of linguistic meaning come
from the beliefs and intentions of the speaker. Section 4.1.1 of the data analysis features a news
article, Indians are black, says Zuma, in which the president chooses a meaning of the word
‘Black’ to suit his political motives at that time. The fact that the president’s understanding of the
semantic attributes of this word was newsworthy shows that his definition was generally not
shared by other South Africans. This is a good example of personalist language ideology in
practice.
“Performative ideology holds that words can have active force; that they can soothe or wound”
(Hill 2008:40) Performative ideologies relate to how language makes people feel. Silverstein
(1979) called ideologies that rationalize and justify the usages and functions of language,
“metapragmatic”. Performative ideology makes it possible to understand some words as harmful,
rather than true or false. Concepts such as ‘hate speech’ and ‘racism’ are understood in terms of
performative ideology.
In addition to their ability to foreground some aspects of discourse and meaning and background
others linguistic ideologies have what Urban (2001 cited in Hill 2008:40) calls a metacultural
function, meaning that they “move culture through the world” and put certain kinds of text and
talk into circulation. Current referentialist linguistic ideology prescribes racist slurs as wrong.
When a prominent public figure uses a racist slur, there is usually great public outcry and an

Christine Cozien
30
ensuing media frenzy. In the attack and defense of this social actor the offending, racist term will
be repeated over and over again, and will appear all over print, electronic, visual and social
media. Although objectors’ intention is to condemn the use of the racist term, often their outcry
inadvertently leads to an increase in its use.
2.1. RACISM
2.1.1 Definition
According to Oxford Dictionaries online, racism is “the belief that all members of each race
possess characteristics, abilities, or qualities specific to that race, especially so as to distinguish it
as inferior or superior to another race or races” and as “prejudice, discrimination, or antagonism
directed against someone of a different race based on the belief that one’s own race is superior”.
Synonyms associated with racism include racial discrimination, racialism, racial prejudice/
bigotry, xenophobia, chauvinism, bigotry, bias, intolerance, anti-Semitism. Of note is that the
only country which receives a special mention in the Oxford Dictionary’s definition of racism is
South Africa, for Apartheid. Yet racism pervades societies all over the world. The recent
Ferguson riots in America are an example of this. Racism did not die with the fall of the
Apartheid regime.
2.1.2 Racism in the South African context
This section discusses the ways in which racism in South African society differs from that of
other societies, and describes some of racial stereotypes and folk theories about race that
currently exist in popular South African culture.
Racism in South Africa differs from racism in most other westernized countries in one
particularly obvious way. All other studies (van Dijk (1991), Hill (2008), Richardson (2007)) of
racism in various westernized countries have emphasised that it is minority ethnic and racial
groups that are the victims of racism, and that the majority group are the perpetrators. In most

Christine Cozien
31
cases it is the minority group who are on the receiving end of racism, but in South African
society racism goes both ways.
During Apartheid it was the minority group who oppressed all other racial and ethnic groups,
especially the majority group. We still suffer the hangover of Apartheid ideology in South Africa
today. Whether we like to admit it or not, many of the racial stereotypes of our past still pervade
our daily societal discourse.
The racist’s guide to the people of South Africa by Simon Kilpatrick provides some examples of
these shared racial stereotypes. The book is intended to be tongue-in-cheek but, in the opinion of
many is in bad taste, and fails as an attempt at good humorous writing. One stereotype is that
Black people steal or are criminals, as seen in this scenario from Kilpatrick (2010: 26) “My
friend’s maid told me that if your neighbour has two chickens and you have none then one of
those chickens is yours. You don’t have to ask if you can have it – it’s yours to take. Whites have
a different philosophy – they prefer to have their own things. This poses a problem for Blacks
because some of them get shocked by electric fences when climbing over the White’s walls to
retrieve their share of the chickens”. The word cloud under section 3.1.3 of the data analysis is
intended to provide evidence of how the media maintains this same stereotype.
The coming of democracy has brought us the traditional sorts of discrimination too. Now that the
majority group is in power we find various examples of racial and ethnic discrimination against
minority groups too. Daily discourse presents us with stereotypes such as, ‘Coloured people are
gangsters’ (Kilpatrick 2010: 116) ‘Indians will always try to rip you off’ (Kilpatrick 2010:90)
and ‘All white people are rich and don’t want to share their wealth’ (Kilpatrick 2010:55-56).
None of these stereotypes are valid, those who subscribe to them simply suffer from the inability
to see others as individuals.

Christine Cozien
32
Many South African’s have a folk theory that ‘Foreigners
2
are all drug dealers/ criminals’. A
2004 study by the Centre for the Study of Violence and Reconciliation found that 87% of police
officers in the Johannesburg area believe that most undocumented immigrants in Johannesburg
are involved in crime, despite there being no statistical evidence to support this view (Masuka,
2006). Xenophobia and xenophobic attacks remain a social problem (viz the July 2014 attacks on
Somali –owned businesses during the Langa service delivery protests, and the looting of shops
owned by immigrants in Soweto in January 2015).
As previously mentioned, much of the racism in South African news media today is covert.
Presupposition (see section 1.1.3.4) has a role here in that much of the content relies on both the
sender and the receiver being acquainted with information about past events, government
policies and local folk theories, such as the stereotypes discussed here. A lot of South African
news discourse and discussions about it on social media make reference to these stereotypes,
policies and events in implicit ways with the help of rhetorical devices like presupposition.
English is the mother tongue of the minority in SA but, because it is the language of economic
prestige and educational advantage, the minority become the privileged group. The majority of
jobs advertised in South Africa today require the candidate to have “an excellent command of
English” or to have “fluent English language skills”. Linguistic discrimination also shows up on
social media where people who comment in one of the other official languages, particularly
Afrikaans, are rebuked with comments like “English please” and “Apartheid is over, speak
English man”.
Recent socio-linguistic research by Mesthrie (2010, 2012) and by Mesthrie, Chevalier and Dunne
(2013) shows that young, middle class South African English speakers of all ethnic groups have
adopted to varying extents phonetic traits ‘once firmly associated with “whiteness”’ (Mesthrie
2
‘Foreigner’ is the South African term for a legal or illegal immigrant from another African, or sometimes Asian,
country. It does not refer to any non-South African national in the traditional sense. The nouns ‘alien’, ‘immigrant’
and ‘illegal immigrant’, which are more commonly used in elsewhere in the world, fell out of favour post the
xenophobic riots in 2008.

Christine Cozien
33
2010:321 citing Lass 1995). Command of the prestige variety is no longer confined to one group
alone, and elite English is becoming, in one sense, deracialised, even if it brings new class
divisions for black people.
3.1 DATA ANALYSIS
I have already discussed the various methodologies I will apply in my data analysis, first briefly
in the introduction, and then in great detail under methodologies of Critical Discourse Analysis.
Therefore I will not repeat myself here. Rather I will refer to and elaborate on the particular parts
of the methodology relevant to the analysis of each individual piece of data as I examine it.
3.1.1 is an analysis of a newspaper article using the methods described from Richardson
(2007)and Fairclough (1989). In 3.1.2 I analyse an image using methods from Understanding
Media (2013), Coppieters (1995) and Kress and van Leeuwen (1990). 3.1.3 looks at word
frequency, visually represented by a word cloud. Here I discuss the possible effects of frequency,
or repetition, on meaning and the generation of stereotypes. 3.1.4 is an example of the folk
theory that racist discourse can be justified by humour. I have investigated the extract as well as
reader response.
Continues overleaf.

Christine Cozien
34
3.1.1 News article – written discourse
Indians are black, says Zuma
April 19 2014 at 11:12am
By Lebogang Seale and staff reporter
Durban - President Jacob Zuma on Friday assured Indians that they were regarded as blacks.
“The Indian community has never been mere auxiliaries in the struggle for freedom, as may be
suggested by some whose sole intention is to divide the black majority for political expediency,”
he told a meeting at Mount Edgecombe in Durban.
Zuma said that while there had been progress in correcting “the injustices of the past meted out
against Africans, Indians, coloureds and women”, representation of black people in top
management positions had only grown from 23.7 percent to 33.2 percent over the last 10 years.
“Much more remains to be done,” he said. “The ANC government will, in the next five years,
relentlessly pursue these policies to ensure that our economy truly reflects the demographics of
our country.”
He called the ruling party the only movement with the commitment and experience to resolve this
issue.
In Johannesburg on Friday, Zuma appealed to religious leaders. “Let us pray for peaceful and
successful general elections on May 7, a right many sacrificed their lives fighting for,” he said,
addressing thousands of Universal Church congregants at Ellis Park Stadium.
As Zuma and his entourage walked into Ellis Park, they welcomed him with applause.
The president sang his trademark song Umshini wam (“bring me my machinegun”) as he
approached the stage.
Figure 1.4 Indians are Black, says Zuma, IOL news, 14 April 2014

Christine Cozien
35
Reflecting on the methods for analysing newspapers, discussed under 1.1.3 Methodology of
Critical Discourse Analysis, three aspects of written discourse appear to be relevant to the piece
of text in figure 1.4, namely foregrounding, referential strategy, and metaphor.
There are a great number of words that may be used to describe an event, which are all accurate
but which do not all carry the same meaning (connoted or denoted; also referential meaning and
indexical or social meaning). By using certain words we choose to foreground certain aspects of
that event to one or another end. Reisgl and Wodak (2001: 45) refer to a text’s ‘referential
strategies’, explaining that opting to describe an event or person in a particular way ‘can serve
many different psychological, social or political purposes’. While the article is indeed a report on
what a social actor said, more accurately it reports what parts of that social actor’s speech the
author of the article chose to foreground, seemingly with the end of creating a story about race
that would grab readers’ attention.
Reading the original speech (Zuma, 2014a) reveals that it consists of discussion on three main
points, delivered in the following order – firstly Easter Friday as a Christian religious
celebration, secondly an explanation of the ANC’s policy on political redress, and thirdly
encouraging South Africans to vote in the upcoming election. The author of this article has
chosen primarily to foreground the second part of Zuma’s speech, the section which focuses on
race, and then has created a sense of looming racial tension during the approaching elections by
attaching the quote “Let us pray for peaceful and successful general elections on May 7, a right
many sacrificed their lives fighting for”. While the president did indeed say all these things, the
point is that he did not say them all at the same time, in the same location, and they were not all
he said. The quote “Let us pray for peaceful and successful general elections on May 7, a right
many sacrificed their lives fighting for” is in fact from Zuma’s speech at Ellis Park,
Johannesburg, on the same day, and not from his speech at Mount Edgecombe (see Zuma
2014b). The selective reporting, or foregrounding, of parts of the president’s speeches from two
separate events fabricates a scene of a threat of racial tension and violence in the run up to the
elections.

Christine Cozien
36
The transition from Durban to Johannesburg is quite subtle, noted only briefly by an introductory
phrase “In Johannesburg on Friday”, in the third last paragraph, and sandwiched in between
quotes from the President’s speeches. I had read this article several times before noticing that not
all of what is reported is verbatim from the president’s speech in Durban, as the headline,
“Indians are Black, says Zuma”, and lead, “Durban - President Jacob Zuma on Friday assured
Indians that they were regarded as blacks”, bring one to believe. Anyone reading the article once
would never have noticed this at all.
This sense of looming pre-election racial tension is augmented through the reporter’s use of
metaphor. The reporter has chosen to use the quote from the president’s Johannesburg speech
because it contains metaphor of war, “sacrificed their lives fighting for”. “Let us pray for peace”
insinuates that there is a possibility of war. Zuma did also call on those gathered at Mount
Edgecombe to pray for peaceful elections, but the words he used “We call on our people this
Easter to pray for peaceful successful elections, the type of elections we have had since 1994”.
These words state the facts – that South Africa has not experienced pre-election racial violence,
and has successfully held peaceful democratic elections, since 1994. They do not provide the
same sense of war and political unrest as his words from Ellis Park do, thus the reporter has
chosen to background them by replacing them with a similar quote from another source that
better suited their rhetorical needs.
Continuing in the vein of metaphor, the use of words such as “entourage” and reporting that the
president sang “bring me my machine gun” strengthen the idea of war and fighting. Watching
YouTube videos of both the Durban and the Johannesburg events
3
shows that the president’s
“entourage” were not his soldiers or bodyguards, but his guests - the Premier of the relevant
region and religious leaders. Umshini Wam, a historical struggle song of the ANC’s military
wing Umkhonto we Sizwe, is particularly controversial because it is reminiscent of the days of
Apartheid and racial civil war in South Africa. The song is often sung by ANC supporters at
ANC rallies. In this case it was sung by supporters at Ellis Park, but not at Durban, which
3
https://www.youtube.com/watch?v=HsGuXMkdDj0 Ellis Park, 18 April 2014
https://www.youtube.com/watch?v=xZPqq88-XsE Mount Edgecombe, 18 April 2014

Christine Cozien
37
provides further explanation for why the reporter chose to tack a report of the Johannesburg
event onto this article about the Durban event – it helps to create a sense of an imminent threat of
racial violence.
In the second paragraph of the article the author arranges parts of Zuma’s speech to create an
ideological square. “Injustices of the past” have been “meted out against” people. Who meted out
the injustices of the past is left unsaid. Most South Africans, via connoted social meaning, will
know that the unknown villain referred to is the Apartheid government. The ANC government,
cast as the hero in this ideological square, will “relentlessly pursue” redress on their behalf. This
us-vs-them, hero versus villain scenario introduces the idea of fighting.
Before moving on from this analysis, two more points deserve clarification. Firstly, the article is
entitled “Indians are Black, says Zuma”. Note the lack of quotation marks in the title, because
the phrase ‘Indians are Black’ was never uttered by Jacob Zuma at either of these two events and
proof of that is that those words appear in neither of his speeches. Secondly, the author translates
the Zulu phrase umshini wam as “bring me my machine gun”, whereas the correct translation is
‘my machine’. It is said that the word umshini in the struggle song Umshini Wam refers to a
machine gun, and being a song of war, it might, but the author’s presentation of “bring me my
machine gun” as the literal translation is incorrect and only exists to provide further metaphor of
war and violence.

Christine Cozien
38
3.1.2 Visual Discourse
Figure 1.5 Photograph of poster at an EFF rally. Source: The News Observer Online, 2013
The poster in figure 1.5 was displayed at an EFF rally in Marikana in November 2013 and
subsequently appeared in numerous South African newspapers (City Press, M&G, Cape Times,
News24, Financial Mail, Argus, Sowetan) was reported on radio and television and littered blogs
and social media online. Very few South Africans were not exposed to this. Some examples of
the headlines that accompanied the image are “Civil war talk from the EFF” (Financial Mail see
Ronbeck 2013) “Malema declares war on whites” (South African News, 2013) “EFF: White
people must learn to share” (Mail&Gaurdian see SAPA, 2013) “Big rise in support for vicious
anti-white party in South Africa” (The News Observer Online, 2013). Note here again all the
metaphors and references to war and violence in relation to race.
The Economic Freedom Front (EEF) is a South African political party which “draws inspiration
from the broad Marxist-Leninist tradition” (EFFighters.org.za, 2014). They consider themselves

Christine Cozien
39
“the vanguard of community and workers’ struggles” (ibid) and say they “will always be on the
side of the people” (ibid). The EFF affiliate themselves with communism by wearing red overalls
and red berets. The overalls show that they represent the worker. The berets are a reference to
Cuba, more accurately Che Guevara, and are a symbol of their revolutionary spirit.
It should be noted that EFF leaders officially distanced themselves and their party from this
racist poster, and made clear that they did not support the sentiment displayed on it (City Press,
2013).
The photograph is taken in such a way that attention is primarily drawn to the text on the sign
and its angry red background. The eye then moves on to the shouting, gesturing, aggressive mass
of all Black people holding the sign, and finally one observes that there are actually more people
behind them, as far as the eye can see. The assumption is that these people are also part of the
demonstration, and that they also support the hateful message on the sign. Cutting off the faces
of those in the background and leaving the partially out-of-focus makes it impossible to
determine whether they are actually part of the rally, or just neutral bystanders, and helps to
create the impression of an endless mass of angry Black people. The photograph is taken from
below, looking up at the angry mass, which makes them appear more powerful and threatening.
The low angle also helps to make the poster (the subject), appear larger than in reality. While the
poster is relatively large, the angle helps to further increase its size, and cutting off a small
section of it, on the right hand side of the photograph, makes it impossible to determine its actual
size, creating the idea of it spanning indefinitely. By positioning the poster so that it almost fills
the entire frame, the photographers creates the sense that it is “taking over”. (How to Analyze a
Photograph, 2013).
The poster is the focal point of the photograph. Analysing the photograph using the rule of thirds
produces the result presented in figure 1.6 on the following page.

Christine Cozien
40
Figure 1.6 Photograph of poster at an EFF rally analysed using the rule of thirds
With rule of thirds the most important, or focal information is presented where two lines cross
(also known as the power point). Here the photographer has positioned two power points over
the message “HONEYMOON IS OVER FOR WHITE PEOPLE IN SOUTH AFRICA”, making
this the focal point of the photograph.
Another power point draws the eye to a raised black fist holding a red spear, a symbol of power,
aggression and violence. The fist is rising up out of South Africa. The final power point draws
attention to the colour red. Kress & van Leeuwen, in their book Reading Images (1990), often
emphasise the importance of colour in images. The colour red is associated with anger, hatred,
violence, blood, war and, often, communism which the Western world has been indoctrinated to
perceive as threatening.

Christine Cozien
41
3.1.3 Word frequency
Figure 1.7: Word cloud generated from Black Males the Face of Crime, Sowetan, 23 December
2010
A word cloud uses the size of the text to represent the frequency of each word in a body of text.
Online applications such Wordle and TagCrowd generate word clouds from bodies of text which
the user inputs. The original article from which the word cloud in figure 1.7 was generated is
appendixed. It was sourced from SA Media, a smaller database of South African print media on
Sabinet.
Sabinet is a large collection of online databases which provides electronic access to news media,
legislative documents and gazettes originally produced only in hard copy. The scanned text is
categorised and tagged with key words to facilitate searching. (The actual text of scans from July
2012 onwards is searchable too because of text recognition in newer .pdf formats).

Christine Cozien
42
Some basic searching within the SA Media database yields some interesting results. Simply
searching the terms black and crime from 27/04/1994 to current date yields 14993 articles,
whereas the terms white and crime yields 11178, and coloured and crime just 1449 results. Why
are black people most often associated with the word crime in SA news media? Another search,
this time associating the words black and poor, white and poor, and coloured and poor, reveals
results of 18506, 13387 and 2036 respectively. Again, why are black South Africans most often
associated with poverty? And how does the frequency with which words are repeatedly
associated affect our perception of different race groups or our construction of racial stereotypes?
A database like SA Media gives an idea of how many individual articles have collocated two or
more specific words, in other words, what the general newspaper-reading public is exposed to in
overall corpus of news media, over several years. But what about word frequency in individual
articles?
Word cloud generators make the analysis of word frequency on more micro level, within one
sample of written discourse, possible and clarify how repetition is used to put a certain message
across. One weakness of the SA Media database search engine is that it is not possible to input
‘NOT’ criteria. So, for example, the word white will often form part of the NP white collar,
which can only be excluded from the white + crime result by manually subtracting the result of
white collar + crime from that of white + crime. The word cloud generator TagCrowd has a
number of features that circumvent these problems.
TagCrowd is an online word cloud generator into which the user can feed text of up to 5MB or
the URL of an entire website. The application will then process the text and present it as a word
cloud (see figure 1.7 above) in which word frequency is visually represented using the size of the
typeface. The more frequently occurring the word, the bigger the typeface. The most frequently
repeated words in a text are what generate the core message.

Christine Cozien
43
It is possible to refine ones results, or impose certain parameters, when using the word cloud
generator TagCrowd. In particular, the user can exclude certain words or phrases, which makes it
possible to cancel out the white versus white collar problem mentioned earlier. In a similar vein
the user can choose whether to group derivatives of the same word or not e.g. learn, learning,
learned.
Additionally the user can set the text language so that the application will ignore common but
irrelevant words, so in English words like a, an, the. TagCrowd also includes the option to show
the word frequency with a word count next to each word in the cloud, and is the only word cloud
generator that has this feature. A minimum frequency may be set, so that if a word in the text less
than X number of times, the generator will exclude it from the results. In order to achieve an
uncluttered, clear result, a maximum frequency can also be set. This means that only words
appearing greater than Y number of times will be displayed in the result, if the user choses to
impose such a parameter. Finally, the conversion of uppercase to lowercase can be controlled by
the user, so that words like PhD will remain that way.
TagCrowd creates a clean, coherent word cloud, which includes numbers for word frequency and
allows the user to control a great number of variables. All this makes it a very useful tool for
linguists investigating word frequency and the power of repetition.
The word cloud in figure 1.5 above is a powerful representation of how the repetition of certain
words has generated the message that Black males are associated with crime. Also, words related
to Black male and crime are police, murders, likely and criminals. The headline Black Males the
Face of Crime suggests that it was the author’s intention to convey this message.
“Used deliberately” says Richard Nordquist, professor emeritus of Rhetoric and English at
Armstrong Atlantic State University “repetition can be an effective rhetorical strategy for
achieving emphasis” According to Fairclough (1989:45) “the effects of media power are
cumulative, working through...repetition...Thus media discourse is able to exercise a pervasive

Christine Cozien
44
and powerful influence in social reproduction because of the very scale of modern mass media
and the extremely high level of exposure of whole populations to relatively homogenous output”
Continues overleaf

Christine Cozien
45
3.1.4 Folk theory – racial stereotyping as humour
Oppikoppi
Oppikoppi is an annual music festival held on a koppie about 200km north of Johannesburg in
the dusty, thorny Limpopo bushveld, outside a town called Northam.
It has been going for 20 years and is now an integral part of white culture.
The festival lasts three or four days, with several stages of music comprising mostly rock, but
also electro, hip-hop and jazz.
Attending Oppikoppi means camping in the bush with about 20 000 other people, all covered in
orange dust and smelling of brandy and sweat.
White people cannot get enough of it. Oppikoppi is like a pilgrimage to reaffirm their white
principles.
Most people spend the entire weekend drunk and unwashed, stumbling from stage to stage,
losing their friends and screaming sporadically into the skies in Afrikaans.
Oppikoppi is the best place to see every decent rock artist in the country in one place.
If one of the 100 bands that plays there later makes it big, you can then say, ‘Oh ja, I saw them
at Oppikoppi,’ even though, at the time they played, you were passed out upside down in the
first-aid tent, waiting to have a nosebleed taken care of.
If you want to understand white people, going to Oppikoppi is a good place to start.
The festival fulfils white people’s love for unpretentiousness, authenticity, Afrikaans, rock music
and camping. It is also a good place to eat boerewors rolls.
More and more black people are starting to attend and perform at the festival, too, so that’s the
transformation objective ticked as well.
Also, something about being drunk in the same jeans for four days makes white people feel like
they themselves are rock stars. And white people like to feel that they are actually unknown rock
stars.
Figure 1.8 - Excerpt from Stuff South African White people like by Christian Lander and Hagen
Engler
Figure 1.8 on the previous page shows an extract from a book by Christian Lander and Hagen
Engler entitled Stuff South African White people like. It is included in the analysis because it

Christine Cozien
46
provides an example of folk theory. In this investigation of folk theory it was the resulting public
discourse on social media that resulted from the publication of this extract that was analysed,
rather than the extract itself.
Observation of public comment and discussions surrounding this publication on the online public
forum, Disqus, as well as on the social media community Facebook, revealed two main
standpoints. One response to this publication in social media was that people felt that the book
and its content were offensive and racist. The screen shot in figure 1.9 below shows a comment
expressing this view, which 119 people agreed with. Yet it also shows that nearly equal number
(102) disagreed with that view point.
Figure 1.9 – Screenshot from the online public forum Disqus of comments about the publication
Stuff South African White people like by Christian Lander and Hagen Engler

Christine Cozien
47
Further analysis of the public comments in figure 1.9 reveals that those forum users who
supported the view that the publication was not racist reasoned this way because they subcribed
to a folk theory whereby racism is excusable because it’s “tongue in cheek” and “because (racial
stereotyping)’s funny”. Figure 1.10 below provides an example of this.
Figure 1.10 Screenshot from the online public forum Disqus showing comments which exemplify
the folk theory that racial stereotyping is excusable if intended humourously.
The comments shown in figures 1.9 and 1.10 represent only a fraction of the comments
discussing the publication, the debate raged on for at least 3 pages and a further 38.5% of the
comments in the public comment stream following echoed this sentiment – that racial
stereotyping and racist discourse are acceptable if done with humorous intent.
Tongue-in-cheek humour has an intriguing semantic aspect to it, in that it dictates which levels
of meaning the receiver may entertain. According to Hill (2008) in circumstances where a
person, public or private, transgresses the linguistic rules regarding what constitutes racist talk,
laid out by referentialist or standard language ideology of that speech community, the blame may
be shouldered in two distinct ways. If the racist talk is believed to be malicious and deliberate
then it referred to as a ‘slur’. However, if there is some desire to defend the utterer or
personalised belief about their intentions being harmless, it is referred to as a ‘gaffe’ or a ‘slip’,
indicating that the utterance was ‘a mistake’ or ‘just a joke’. Within this framework societal

Christine Cozien
48
members are able to assign good or bad moral attributes to a particular person or social actor. By
this logic racist, discourse may be excused by virtue of its intention to be humorous or ‘tongue in
cheek’, i.e. not serious.
The question then is why nearly 50 percent of the participants in the forum discussion chose to
defend these authors work as a ‘gaffe’. One might reason that it is because they are racist, but
this line of thinking seems a little pedestrian. A more plausible hypothesis then is that a folk
theory exists among the South African English speaking community that racist talk and racial
stereotyping is excusable if it is done in so-called ‘good humour’. The racist’s guide to the
people of South Africa by Simon Kilpatrick, and the comedy show inspired by it – Race card by
Siv Ngesi, are examples of written and spoken discourse which provide evidence for the
existence of this folk theory. The book seems to have been very popular with the English
speaking South African public – it has been reprinted eleven times, as many as five times in a
year. Its reviews reveal that many South Africans do seem to believe that racist discourse and
racial stereotyping is acceptable, even more, it is funny, if it is imparted under the banner of
comedy or humour. Die Rapport newspaper say they were reduced to “tears of laughter” while
reading The racist’s guide to the people of South Africa, and The Weekend News describe the
book’s “humour” as “side-splitting”. Further data collection and analysis of South African text
discourse similar to Stuff South African White people like and The racist’s guide to the people of
South Africa, as well as of spoken discourse presented at South African comedy shows like Race
card, would be necessary to draw any definite conclusions as to the existence of this folk theory.
It could be argued that this sort of humour has therapeutic value, in that it exposes private
negative stereotypes and encourages South Africans to evaluate the validity of these stereotypes.
During Apartheid racial humour and satire was discouraged. Perhaps as a result, in the new
South Africa satire (particularly cartoons) has become means discussing sensitive issues
surrounding race and politics, with the medium of humour acting as a proverbial safety net.
Humour, it would appear, enables people to discuss issues that would otherwise be taboo by
virtue of the understanding that it is not to be taken too seriously.

Christine Cozien
49
CONCLUSION
This paper has taken an exploratory look at what the analysis of discourse in the South African
media can reveal about racism and racial stereotyping. Four different methods of critical
discourse analysis were used, with the aim of determining which ones produced the most
informative results. Richardson’s (2007) methodology for analysing newspapers were applied to
the analysis in section 3.1.1 and produced revealing results, albeit when supplemented by
additional research to obtain verbatim records of the speeches reported on in the news article.
The word cloud in section 3.1.2 is a powerful visual representation of word frequency and shows
potential as a tool for discourse analysis. The analysis of visual discourse in 3.1.3 using methods
from and Kress and van Leeuwen’s Analysing images (1990) and Coppieters (1995) was difficult
and results do not feel particularly conclusive or enlightening. The analysis of a particular genre
of written and spoken discourse, namely humour, as well as from public online forum
discussions surrounding it in 3.1.4 has drawn some initial conclusions about folk theory, but
requires that many more similar pieces of discourse be analysed before any solid conclusions can
be drawn. A paper on discourse analysis which arrives at any relatively firm conclusions is, by
virtue of the data, relatively long. This is so because one most constantly extract and quote parts
of the discourse during the analysis. This paper’s length placed limitations on the amount of data
that could be analysed, and thus was better used to investigate possible methods of analysis for
future studies.
Considering the above, it seems that from this point the most appropriate course of action would
be to embark on three separate research projects. One analysing text discourse from South
African news media using Richardson’s (2007) methodology, investigating the role of written
discourse, from lexeme to narrative, in promoting racism. Another analysing text discourse using
word clouds and investigating the role of word frequency in the generation of racial stereotypes.
And finally a third researching folk theory surrounding racial stereotyping, via the analysis of
written and spoken discourse under the genre of humour. Racism has many facets. Several
different approaches, more thoroughly applied have the potential to produce new insights into the
nature of racial stereotyping, and in turn, what that reveals about the relationship between
language and culture.

Christine Cozien
50
An accurate and detailed understanding of the construction of racial stereotypes and implicit
biases in society demands a multidisciplinary approach to research. This paper has briefly
engaged works from the disciplines of philosophy, anthropology, linguistics and media and
communication. In future expansions of research on this topic I should like to also engage fields
such as sociology, specifically I would like to engage two test scales, one originating from
sociological research and the other from psychological research. The first would be based on
Williams’ (1997) Everyday Discrimination Scale, tailored to be relevant in the South African
context. The second would be word association test, inspired by Jung’s psychological work on
unearthing the contents of the human subconscious. The word association test links in with a
hypothesised relationship between collocation and cognitive association. I hypothesise that the
lexemes most frequently collocated with racial category lexemes should in turn be associated in
the subconscious, and word association should determine whether this is the case or not.
Collocations are “familiar groupings of words...that habitually appear together and thereby
convey meaning by association” (Nordquist, nd). I hypothesise that the more often a word is
collocated with another, the more likely it is to be attributed as a semantic property of its
collocated word. In the famous words of British linguist, John Firth, “You shall know a word by
the company it keeps”.
As emphasised by van Dijk’s (2004) assertion that it should rather be called Critical Discourse
Studies, CDA stresses the importance of an interdisciplinary approach to the analysis of texts.
CDA supports Halliday’s point of view that we should reject the idea of language being
something that can be studied in isolation. Language is inextricably and intimately link to social,
cognitive, physical and biological aspects of life. This view is an expansion of Fairclough and
Wodak’s (1997:276) sixth principle that “relationships between text and society are mediated
and a socio-cognitive approach is needed to understand how these link”.
As Mesthrie (2012:371) quite accurately points out, “race and color are fuzzy concepts, not at all
biologically determined (but) dependent upon power and cultural relations within a territory”.
The lexemes used to describe so-called racial categories change, as do their meanings, and it is
usually the political power of the time which determines this. Politicians often excuse the use of

Christine Cozien
51
racial categories by claiming that they are necessary to redress the economic and educational
inequalities of the past. One might counter that politicians abuse these labels to stir up voters, or
cover up their own short comings by diverting attention, or laying blame, elsewhere. A careful
analysis should look at what contexts we are still using these labels in, and in which of those
contexts could we afford not to use them, without sacrificing any progress we have already made
towards addressing the wrongs of the past and achieving equality.
It seems also that many people, not just in South Africa but in other countries too, think it is
acceptable to use racial category lexemes to describe people, yet this preserves the idea that race
is a real thing. That race is real is in itself wholly debatable. However one thing is relatively
certain – that the use of racial category lexemes, whether malicious or not, perpetuates the idea
of differentness, and is contrary to any progress towards a non-racial society.

Christine Cozien
52
REFERENCES
Ahearn, Laura. 2012. Living language: An introduction to linguistic anthropology. Vol. 2.
Malden, MA: Wiley-Blackwell.
Braude, Claudia. 1991. Cultural bloodstains: Towards understanding the legacy of apartheid
and the perpetuation of racial stereotypes in the contemporary South African media.
Human Rights Commission, Interim Report into Racism in the Media. Johannesburg:
South African Human Rights Commission.
Breeze, Ruth. 2011. Critical Discourse Analysis and its critics. Pragmatics 21:4.493-525
Brown, Gillian and George Yule. 1983. Discourse Analysis. Cambridge: Cambridge University
Press.
City Press. 2013. EFF disappointed over ‘honeymoon is over for whites’ banner. City Press, 15
October 2013. Accessed from http://www.citypress.co.za/politics/eff-disappointed-
honeymoon-whites-racist-banners/ [03 January 2015]
Carroll, John. (ed.) 1956. Language, Thought, and Reality: Selected Writings of Benjamin Lee
Whorf. Massachusettes: MIT Press.
Coppieters, Kris. 1995. Analysing a Photograph – a how-to guide. University of Oregon.
Available from http://pages.uoregon.edu/jlesage/Juliafolder/PHOTOANALYSIS.HTML
[24 September 2014]
Crystal, David. 2008. A Dictionary of Linguistics and Phonetics. 6th Edition. Oxford: Blackwell
Duranti, Alessandro. 2001. Linguistic Anthropology: History, Ideas and Issues. In Alessandro
Duranti (ed.) Linguistic Anthropology, pp 1–38. Oxford: Blackwell.
Fairclough, N. 1995. Critical Discourse Analysis: The Critical study of Language. London:
Longman.
Fairclough, Norman. 1989. Language and Power. 2nd ed. Essex: Pearson Education Limited.

Christine Cozien
53
Human Nature: Justice versus Power – Noam Chomsky debates with Michel Foucault. 1971a.
Interview by Fons Elders as part of the International Philosphers’ Project debate series.
Sourced from http://www.chomsky.info/debates/1971xxxx.htm. [10 July 2014]
Foucault - The lost interview. 1971b. Video footage by Fons Elders, posted on Youtube by
Lionel Claris. Sourced from https://www.youtube.com/watch?v=qzoOhhh4aJg. [10 July
2014]
Fultner, Barbara. 2011. Jurgen Habermas: Key Concepts. Durham: Acumen.
Halliday, Michael. 2006. Linguistic studies of text and discourse. Vol. 2. London: Bloomsbury.
Habermas, Jurgen. 1988. On the Logic of the Social Sciences. Translated by Shierry Weber
Nicholson and Jerry A. Stark. Michigan: The MIT Press
Habermas, Jurgen. 2008. Between Naturalism and Religion: Philosophical Essays. Cambridge:
Polity Press.
Hill, Jane. 2008. The everyday language of White racism. Massachusetts: Wiley-Blackwell.
How to analyze a photograph, 2013. Available from
http://www.understandmedia.com/topics/media-theory/111-how-to-analyze-a-photograph
[24 September 2014]
Hoskovec, Tomas. 2012. Perspective fonctionelle de la phrase – l’apport de Cercle du Praque.
Revue semestrielle de linguistique et litteratures romanes. 8:1, 7-25.
Hymes, Dell. 1962. The Ethnography of Speaking. In T. Gladwin & W. C. Sturtevant (Eds.),
Anthropology and Human Behavior. pp. 13–53. Washington, DC: Anthropology Society
of Washington.
Irvine, Judith. 1989. When talk isn’t cheap: Language and political economy. American
Ethnologist. 16:? 248-267.
Janks, Hilary. 1997. Critical Discourse Analysis as a Research Tool, Discourse: Studies in the
Cultural Politics of Education, 18:3, 329-342
Kress, Gunther & Theo van Leeuwen. 2006. Reading Images: The grammar of Visual Design.
London: Routledge.

Christine Cozien
54
Luke, A. 2002. Beyond science and ideological critique: Developments in critical discourse
analysis. Annual Review of Applied Linguistics 22: 96-110.
Malema declares war on whites. SouthAfricanNew.info, 14 October 2013. Accessed from
http://www.southafricannews.info/Article.php?ID=444343& [25 September 2014]
Malle, Betream. 2006. Folk Theory of Mind: Conceptual Foundations of Social Cognition. In R
Hassin, J Uleman & J Bargh (Eds.) The New Unconscious (Social Cognition and Social
Neuroscience). pp 1-29. Oxford: Oxford University Press.
Mangena. Isaac. 2007. ‘Umshini Wam echoes through SA’. The Mail & Guardian, 23
December. Available from http://mg.co.za/article/2007-12-23-umshini-wami-echoes-
through-sa [25 September 2014]
Martin, J. 2004. Positive discourse analysis: Solidarity and change. Revista Canaria de Estudios
Ingleses. 49:?,179-202.
Masuka, Themba. 2006. Targeting Foreigners: Xenophobia among Johannesburg’s police.
Institute for Security Studies. Crime Quarterly No 15. Retrieved 6 September 2008.
Mesthrie, Rajend. 2010. Socio-phonetics and social change: Deracialisation of the GOOSE
vowel in South African English. Journal of Sociolinguistics 14(1): 3–33.
Mesthrie, Rajend. 2012. Social change and changing accents in South Africa. In Philip
Seargeant, & Joan Swann (eds.) English in the World: History, Diversity, Change pp
316-322. Abingdon, England: Routledge
Mesthrie, Rajend. 2012. Race, Ethnicity, Religion, and Castes. In J.M. Hernandez-Campoy &
E.J. Camilo Conde-Silvestre. Handbook of Historical Sociolinguistics. Oxford: Wiley-
Blackwell.
Mesthrie, Rajend, Alida Chevalier & Timothy Dunne. 2013. A Study of Variation in the BATH
vowel among White Speakers of South African English in Five Cities. University of
Pennsylvania Working Papers in Linguistics. 19 (20: 131-140.
Nordquist, Richard. nd. Collocation (Words) blog post on grammar.about.com Sourced from
http://grammar.about.com/od/c/g/collocationterm.htm [13 July 2014]

Christine Cozien
55
Nordquist, Richard. nd. Repetition (Language and Rhetoric) blog post on grammar.about.com
Sourced from http://grammar.about.com/od/c/g/collocationterm.htm [9 September 2014]
Nursya’bani, Fira et al. 2013. Critical Discourse Analysis. Slide Share presentation. Indonesia
University of Education. Accessed from http://www.slideshare.net/FirraBannie/critical-
discourse-analysis-17133325. [24 January 2015]
Oxford Dictionaries. nd. Online edition. Oxford: Oxford University Press. Accessed from
http://www.oxforddictionaries.com/ [25 January 2015]
Reah, Danuta. 2002. The Language of Newspapers. 2nd edition. New York: Routledge
Reisigl, Martin and Ruth Wodak. 2001. Discourse and Discrimination: Rhetorics of racism and
antisemitism. New York: Routledge
Richardson, John E. 2007. Analysing Newspapers – An approach from Critical Discourse
Analysis. Hampshire: Palgrave Macmillan
Ronbeck, Phillip. 2013. ‘Civil war talk from the EFF’ Financial Mail, 24 October. Available
from http://www.financialmail.co.za/opinion/letters/2013/10/24/letter-civil-war-talk-
from-the-eff [25 September 2014]
SAPA, 2013. ‘EFF: White people must learn to share’ Mail &Gaurdian, 15 October. Accessed
from http://mg.co.za/article/2013-10-15-eff-white-people-must-learn-to-share [25
September 2014]
Sapir, Edward. 1921. Language. New York: Harcourt Brace.
Sanpath, Arthi. 2014. ‘Now it’s water shedding’, The Sunday Independent, 13 December.
Accessed from http://www.iol.co.za/ios/news/now-it-s-water-shedding-
1.1795227#.VI1O0To9LIU [14 December 2014]
Seale, Lebogang. 2014. ‘Indians are Black, says Zuma’ IOL news, 19 April 2014. Accessed from
http://www.iol.co.za/news/politics/indians-are-black-says-zuma-
1.1678029#.VMXScTo9LIU [25 January 2015]
Schegloff, E. 1997. Whose text? Whose context? Discourse and Society 8.2: 165-187.
Schiffrin, Deborah. 1994. Approaches to Discourse. Oxford: Blackwell.

Christine Cozien
56
Silverstein, Michael. 1979. Language structure and linguistic ideology, in Paul Clyne, William
Hanks and Carol Hofbauer eds. The Elements: A parasession on Linguistic Units and
Levels. Chicago: Chicago Linguistic Society
Simpson, P. 1993. Language, Ideology and Point of View. London: Routledge.
Stubbs, Michael. 1996. Text and Corpus Analysis. Oxford: Blackwell.
The News Observer Online. 2013. ‘Hard Times ahead: Big rise in support for vicious anti-white
party in South Africa’ 31 October. Accessed from http://newobserveronline.com/hard-
times-ahead-big-rise-in-support-for-vicious-anti-white-party-in-south-africa/ [25
September 2014]
Toolan, M. 1997. What is critical discourse analysis and why are people saying such terrible
things about it? Language and literature 6.2: 83-102.
van Dijk, Teuns. 1984. Prejudice in Discourse: An Analysis of Ethnic Prejudice in Cognition
and Conversation. Amsterdam: J. Benjamins Co.
van Dijk, Teuns. 1985. Discourse and Communication: New Approaches to the Analysis of Mass
Media Discourse and Communication. Walter de Gruyter: Berlin.
van Dijk, Teuns. 1987. Communicating Racism: Ethnic Prejudice in Thought and Talk. Newbury
Park, CA: Sage Publications.
van Dijk, Teuns. 1991. Racism and the Press. London: Routledge.
van Dijk, Teuns. 1993. Elite Discourse and Racism. Newbury Park, CA: Sage Publications.
van Noppen, Jean-Pierre. 2004. CDA: A discipline come of age? Journal of Sociolinguistics 8:1,
107-126.
Vygotsky, Lev. 1978. Mind in Society. London: Harvard University Press
Whorf, Benjamin-Lee. 1939. The Relation of Habitual Thought and Behaviour to Language, in
Carroll, Language, Thought, and Reality: Selected Writings of Benjamin Lee Whorf.
Massachusettes: MIT Press.
Widdowson, H. 1998. The theory and practice of Critical Discourse Analysis. Applied
Linguistics 19.1:136-151.

Christine Cozien
57
YourDictionary.com. 2014. Click-bait. Accessed from http://www.yourdictionary.com/click-bait
[December 2014]
Zuma, Jacob. 2014a. Address by President Jacob Zuma to the Indian Community. 18 April 2014.
Mount Edgecombe, Durban, KwaZulu-Natal, South Africa. Accessed from
http://www.anc.org.za/show.php?id=10866 [03 January 2015]
Zuma, Jacob. 2014b. Address by His Excellency, President Jacob Zuma, during the Easter
Weekend Service of the Universal Church of the Kingdom of God. 18 April 2014. Ellis
Park, Johannesburg, Gauteng, South Africa. Accessed from
http://www.anc.org.za/show.php?id=10865 [03 January 2015]