The Impact of a Monolingual Medium of Instruction in a Multilingual University in South Africa

Abstract This paper investigates the impacts of a monolingual medium of instruction (MOI) in a multilingual university in South Africa. It offers a critical examination of the different effects this may have on learning and academic performance of the students. It further examines why students may or may not prefer the selected medium of instruction and the effect this has on their learning abilities. Hence the analysis of these different views would inform and improve the teaching and learning in classrooms at higher education level. In addition, the paper presents factors that appear to play a role in shaping the students’ ideologies and attitudes towards English, a language used as MOI across most universities in South Africa and which many of the students appear to find difficult to use in academic learning. The findings from this qualitative study are examined using studies on language ideologies, policy and planning.


INTRODUCTION
South Africa is a multilingual country with an extraordinary diversity of races, groupings, languages and landscapes that characterises this country. It has the most diverse and accommodating language policy in the world (Bamgbose 2003). South Africa has eleven official languages, and a significant number of minority languages. Despite the large number of official languages, English is the dominant language in South Africa (Dyers 2008) and as a result, many parents do their best to ensure that their children are educated in English. Besides the eleven official languages, there are a number of other African languages spoken in SA. Other SA languages covered by the constitution are: Khoi, Nama and San languages, South African Sign Language (SASL), Arabic, German, Greek, Gujarati, Hebrew, Hindi, Portuguese, Sanskrit, Tamil, Telegu and Urdu. There is also one pidgin: Fanagalo, which was used largely in the mining industry (Allmann 2009), but now appears to be in decline owing to the negative sentiments around its continued use. South Africa also has different varieties of English influenced by social class, level of education and region. For example, in the city of Cape Town we find Cape Flats English, spoken in Cape Town and its surrounding areas (McCormick 2000). There is also South African Indian English (Mesthrie 1992), spoken mainly, but not exclusively, in the province of Kwa-Zulu Natal, and Black South African English (BSAE) (Da Silva 2008). English is also used in many merged or blend-ed varieties like Tsotsitaal and Hip-hop language (Williams 2012).
SA therefore has no single culture as a result of this diversity, and the new SA is thus marked by multiracialism and multiculturalism. Although SA has eleven official languages, English is understood across the country and is the language of media, politics, business and serves as the lingua franca of this country (Beukes 2003).

An Overview of the Language Policy in
South Africa (SA) Beukes (2008) provides a very comprehensive overview of the developments of language policy in South Africa. The colonial era, dating from the settlement of the Dutch in SA in 1652 to the period of British rule and the establishment of the Union of SA, showed scant respect for the indigenous languages of South Africa. During the apartheid era, when the country was ruled by the Afrikaner Nationalists, the indigenous languages became part and parcel of a strategy to divide and rule the black majority in the country. The Afrikaner nationalists, obsessed with racial stratification, emphasized what they saw as the links between ethnicity and language, especially in their creation of small separate 'homelands' for each black ethnic group, like Transkei for the Xhosas, Zululand for the Zulus and Bophuthatswana for the Tswanas. Separate language boards were created for each indigenous language, and quite a lot of work was done on developing dictionaries and literature in these languages, which were used in especially prima-ry education and other homeland domains. (Standard) Afrikaans and English were the only official languages in this period, and other varieties of these languages, for example, Cape Flats Afrikaans, Black South African English or South African Indian English, had little or no status apart from being markers of 'Coloured', 'Black' or 'Indian' identity.
According to Beukes (2008) a serious attempt was made to address the subordinate situation of the indigenous languages when the country became a democracy in 1994. In order to promote linguistic diversity in South Africa (SA), the Department of Arts and Culture organized a national conference on language, which led to the establishment of the Pan South African Language Board (PanSALB).This Board was formed to promote broader multilingualism, aid the development and wider use of all the eleven official languages as well as the Khoi, Nama, and San languages and the South African Sign Language. The national Department of Education developed the Language in Education Policy (LiEDP) that stressed multilingualism to promote diversity and build a non-racial SA.
The language policy in SA as said earlier is regarded as the most accommodating policy in the world (Bamgbose 2003) as it gives room for 11 official languages, of which nine are African languages. The two official documents that constitute the framework for language planning in SA are the Constitution and the SA Language Bill (The South African Constitution 1996 No. 108.6).
However, since the inception of the two documents, the SA National Language Policy has been criticized for not fulfilling its lofty aims (Webb 2002;Beukes 2008). Although the policy is accompanied by a clear Implementation Plan, progress has been slow. English, the default lingua franca of this country, has in fact become the dominant language of the country, particularly as it is favoured by politicians and businesses (Abongdia 2013).
The main SA languages, according to Webb (2002), are rooted in the political history of the country. He contends that English today is a very prestigious language, followed by Afrikaans to a lesser degree in certain parts of the country, while the indigenous languages have very little educational and economic value. For many speakers of these languages, according to Webb, the indigenous languages are seen as symbols of "uneducated, traditional, rural, culturally backward people with lower mental power" (Webb 2002: 8). Paradoxically, this is not the case with Afrikaans as it is the only language that is competing with English in SA (Kamwangamalu 2000;Abongdia 2013) at the official level. Webb views English and the Bantu languages as holding opposite positions and concludes that English can be used for discrimination and manipulation, as it is a vehicle for the struggle of power between the different socio-economic groups (Webb 2002: 8). He therefore suggests a language policy that should be developed in a way that it can facilitate rather than impede development in all the domains.
Beukes (2008) is of the opinion that the government lacks a commitment to its own policy because language matters are being relegated to the background. She argues like Pennycook (2000) and Ricento (2000) that language policy and planning are contextually determined and for it to be implemented, note should be taken of the social, cultural and economic interest benefits to its citizens. She goes further to say that traditionally the language policy activities followed the top-down approach model where people in authority take decisions for groups with little or no consultation with those who use the language.
However, it should be noted that the use of local languages in formal education has been highly problematic in sub-Saharan Africa (Abongdia 2013). Mother-tongue instruction is permitted in primary grades, but the implementation of this policy is not yet common in many African countries, with the exception of Malawi, Zimbabwe, Tanzania, Mozambique, Cameroon, Mozambique and South Africa. In Cameroon, there is no clear cut rule to guide the implementation of the mother tongue medium of instruction. However, it runs across the first three years of schooling (Trudell 2007). In the case of SA the indigenous languages are used in the first three years of schooling and later used/ learned as a school subject (Stroud 2004). One of the main reasons Trudell (2007: 561) puts forward for this practice is "…the perspectives and desires of the stakeholders in the education process who reside in the local community: teachers, parents, community leaders, school authorities, and children themselves. Responding to their particular historical, social, and economic context, these stakeholders have developed their own perspectives on the nature and outcome of formal education as well as beliefs about the utility and proper place of the local language." This means that mother-tongue instruction is not used in all areas in the country, but mostly in local communities depending on the agreement of the stakeholders residing in these communities/localities.
Most national policy makers and community level educators often hesitate to replace education in official languages with education in local languages. Trudell (2007: 553) also argues that the national level of adaptation of mothertongue education is not only impeded by pedagogical aspects but also by political and social factors that have a powerful role when language and educational issues are under consideration. On the other hand, the implementation of mother-tongue education in the African community contexts is more of a local endeavour and is greatly influenced by the values and beliefs of those community members (Abongdia 2013).

Language Policy of the University of the Western Cape
The University of the Western Cape is the focal points of this study. This university was established by an Act of Parliament as an ethnic college for "coloured" students in 1959. It was opened as a University College (mainly aimed at training high school teachers) in 1960, raised to the status of a full university in 1970 and gained its autonomy in 1983 through the University of the Western Cape Act. In 1982, it began the process of transforming itself into an internationally recognised university with a reputation for excellence in teaching and research. Due to its experience in the liberation struggle, UWC is aware of a distinctive academic role in helping to build an equitable and dynamic society. Today it has a diverse student and staff population.
The 2003 language policy at this university aims at "ensuring equity, social development and a respect for South Africa's multilingual heritage" (UWC Language Policy 2003). The language policy "attempts to guide institutional language practice so that it furthers equality, social development, and a respect for our multilingual heritage" (Language Policy Preamble 2003). The policy outlines which languages will be used by the institution in terms of language of teaching and learning, assessment, access to academic and professional discourse, promoting multiculturalism, and language of internal and external communication. The main language used at the institution is English, followed by Afrikaans and isiXhosa. The choices of these languages were guided by the fact that these are also the official languages of the Western Cape Province. It is logical that tertiary institutions should promote the most used languages in their provinces.

Objective of the Study
Given the nature of the language policy of the university under study and its diverse population, the paper seeks to investigate the impact of a monolingual medium of instruction in this multilingual and why not multiracial institution.

Theoretical and Conceptual Framework
Theoretically and conceptually, this paper interrogates the notions of language policy, language ideology, and medium of instruction within the context of the late-modern era and globalization. Blackledge (2005:32) contends that language ideologies "are positioned in, and subject to, their social, political and historical contexts", and are inscribed in many "chains of discourse" (ibid: 209). Language policy is one of these discursive chains, and cannot be divested from the politics and history of either the country or institution which draws up such policies (Abongdia 2013;Dyers and Abongdia 2014). Given their centrality to this paper, the writer commence with an examination of some of the definitions of the concepts language ideologies and language policy. Irvine (1989:255) defines ideologies as "the cultural system of ideas about social and linguistic relationships, together with the loading of moral and political interests". This definition emphasizes the 'moral and political interests' that underpin societal ideologies, and shows that ideologies cannot exist outside of the power relationships in any society. This is further emphasized in Fairclough's definition of ideologies (2003:9) as: "representations of aspects of the world which contribute to establishing and maintaining relations of power, domination and exploitation. They may be enacted in ways of interaction (and therefore in genres) and inculcated in ways of being identities (and therefore styles)".
It is the contention of this paper that ideologies about language are an integral part of ideologies in other domains of human activity such as the creation of language policy. Irvine and Gal (2005:35) see language ideologies as "the ideas with which participants and observers frame their understanding of linguistic varieties and mapped such understandings onto people, events and activities that were significant to them. These ideologies are not only held by the immediate participants in the local sociolinguistic system but also by observers like linguists and ethnographers, who have put in place, boundaries between people and languages." This definition emphasizes how ordinary people understand the role of different language varieties, and link this understanding to their perceptions of, and responses to, the speakers of such varieties. In addition, the definition also shows how people perceive and interpret the role of language in major events that affect their lives (for example, political speeches, dealing with government departments and other state institutions like hospitals) as well as the ways in which particular language varieties impact on their daily activities (for example,at work, in schools and universities and places of worship).
All the above definitions are framed by Kroskrity's (2000: 8-21) four intersecting dimensions of language ideologies. According to him (his italics): "…language ideologies represent the perception of language and discourse that is constructed in the interest of a specific social or cultural group"; "…language ideologies are profitably conceived as multiple because of the multiplicity of meaningful social divisions (class, gender, clan, elites, generations, and so on) within sociocultural groups that have the potential to produce divergent perspectives expressed as indices of group membership"; "…members may display varying degrees of awareness of local language ideologies"; and lastly "…members' language ideologies mediate between social structures and forms of talk." The first dimension reveals how people's notions about languages are rooted in their social experiences and often tied to their political and economic interests. This dimension comes out strongly in studies conducted by Lippi-Green (1997), Cameron (1995) and Milroy and Milroy (1998) on the type of language-based discrimination that takes place in the United States and Britain against marginalized social groups who do not speak the 'standard' language. In the setting of South Africa English dominates the other official languages. In view of this, Blommaert (2006) argues that language ideologies are not attributed to one person or located in a particular site (political party/government) but it penetrates the whole fabric of societies or communities and results in normalised, naturalised patterns of thought and behaviour. Thus ideology is seen as common sense, the naturalised activities that sustain social relation and power structures and the patterns of power that reinforce such common sense (Bourdieu 1991). Kroskrity's (2000) second dimension reveals that social experiences are not uniformly distributed but differ in terms of the divisions in particular societies. Men and women, older and younger people, for example, may have very different perceptions, beliefs and attitudes about language. In this way, different ideologies can become contentious within the same group and lead to tensions between what are often stateendorsed dominant ideologies and their opponents. In South Africa, for example, we find contrasting views on the perceived shift towards English from Afrikaans and Xhosa depending on the area in which respondents live and the particular social class they belong to (Dyers 2008a, b;Anthonissen and George 2003).
At the core of the third dimension is how conscious members of a society are about their attitudes towards languages. Kroskrity contends that those who are most conscious of their ideologies are the ones likely to be most vocal about the value of different languages. But ordinary people's ideologies are more likely to be reflected in their actual language usage -the languages they prefer to use as opposed to the languages they avoid using (Kroskrity 1999: 19;Dyers 2008b;Dyers and Abongdia 2014).
In his fourth dimension, he shows how people use their ideologies about language as a bridge between their 'sociocultural experience and their linguistic and discursive forms as indexically tied to features of their socio-cultural experience.' In other words, people are quite selective about the features of language and the role of particular languages in society when expressing their language ideologies. Certain features stand out, for example "I don't like the sound of that language/language variety -it sounds too coarse/common/disrespectful" or "My language cannot be used at university level". In the first example quoted here, the language or language variety has somehow become associated with unpleasant social experiences and in the second example there is recognition of the absence of 'my language' in a tertiary setting. A good example of this dimension is Swigart's analysis of the harsh criticism by the Senegalese Francophone elite of former President Abdou Diouf's speech to the nation in Urban Wolof, one of the local Senegalese languages, instead of French (Swigart 2001).
Ideologies about language are of course not about language alone, but reflect issues of social and personal identity (which includes social status), and are reflected in actual linguistic practice -how people talk, what they say about languages, the language choices they make for themselves and their children, and what they regards as important and essential languages for both survival and advancement in terms of employment opportunities and social standing. These issues of social and personal identity are reflected in Weber and Horner's (2012:16) definition of language ideology as the "cultural system of idea and feelings, norms and values, which inform the way people think about language". Of relevance to this paper are the five major language ideologies listed by Weber and Horner (2012: 16-20).
There is a bilateral relationship between language ideologies and attitudes on the one hand, and language planning and policy on the other. Planning languages and writing language policies are strongly influenced by the prevailing language ideologies and possibly more personal attitudes of those who are directed by governments to do this kind of work (Abongdia 2013;Dyers and Abongdia 2014). At the same time, she argues that language planning and subsequent policies put in place can also influence people's language ideologies and attitudes as well as their social practices.
Language planning is defined by Kaplan and Baldauf (1997: 3) as "a body of idea, laws, and regulations (language policy), change, rules, beliefs, and practices intended to achieve a planned change (or to stop change from happening) in the language use in one or more communities. To put it differently, language planning involves deliberate, although not always overt, future oriented change in systems of language code and/or speaking in a societal context". According to Mühlausler (2000), languages are not isolated systems but interact with other systems outside linguistics such as culture, politics, and environment. Spolsky and Shohamy (2000) define language policy as "an effort by someone with or claiming authority to change the language practice (or ideology) of someone else". Within the language policy, the policy maker has some level of authority over those expected to follow its requirements. They make reference to Ager (1996) who sees language policy as obtaining power rather than distributing it. When creating a language policy, the question to be answered is: for whom is this policy intended? And who are the policy makers? Spolsky and Shohamy think that if language policy aims at changing language practice, then there is a need to study not just policy making but the implementation and evaluation of the policy as well. For policy to be analysed, it becomes very important to evaluate its impacts. It is important to note that sometimes the statement of the policy can be more important than its effective implementation. Brumfit (2006) in his contribution to language policy says that all humans are born with the ability or capacity to acquire a language(s) that they are exposed to. They have no genetic links to a particular language. He further notes that thresholds are found both between and within languages with liminality as a characteristic of contemporary language use. Liminality is a situation where people shift from (and eventually lose) their original identity, role and positions. This happens as the liminal person moves from one place to another dropping and acquiring new values (Pavlenko and Blackledge 2004;Turner 1982). Turner (1982) further says that languages are constructed just like identity depending on the political, individual or group needs/choices. The name of a language depends on the particular social group, thus it should be seen as a social act. Many of the world's languages have the names of particular ethnic groups they are associated with, for example,isiXhosa as the language of the Xhosa speech community in South Africa, isiZulu, Sepedi, Sesotho, Setswana, siSwati etc. Brumfit (2006:36) in his research on language policy in Europe argues that "policy making is normally aspirational and symbolic". What is significant about a policy is what is excluded more than what is included, and what is considered desirable to the policy makers. Conversely, Brumfit thinks that a policy might run the risk of not being practical and not meeting the needs of the people. "People's individual repertoires are modified, changed and widened because of language contact as a result of the permeability of language boundaries, and such extended repertoires frequently find no reflection in language policies" (Brumfit 2006: 37). The wish to extend these repertoires is based on rules put in place by those of the language(s) that one wishes to add to one's repertoire. Such extended repertoires can be seen in people's use of the many cross-border languages of Africa as well as the heteroglossia (Bakhtin 1981) so prevalent in its multilingual cities. Brumfit (2006: 39) further sees multilingualism and multi-dialectalism as having always been the norm for most people, and views policy as a way of controlling or managing linguistic diversity. In this regard, he refers to English as the default language for international communication but acknowledges the need to provide support for minority languages. Brumfit is critical of monolingual, monocultural speakers of English, and argues that they should be made to see foreign languages as means of expanding their repertoires. Thus, language policy for him should help to "reduce the negative effects of people hiding behind a language if it acknowledges the permeability of languages in all directions" (Brumfit 2006: 40). Language is "the Epiphenomenon in power relations" that can easily be used to define group and individual identities (Brumfit 2006: 41), and he sees the naming of languages as political and not linguistic in nature, given that individuals speaking the same language use different lexicons, syntactic features, cultural codes, accent, as well as tone of voice. In addition, language rights are often promoted over individual rights.
However, while policies may ostensibly suggest that they are instruments of language rights, as in the case of SA, individuals generally do not have the right to use the language they are most comfortable in (irrespective of its official status) but have dominant languages of power imposed on them. Given this reality, Brumfit (2006) concludes that policy makers should make it clear that language policies are political, rather than linguistic instruments. He further argues that it is not easy to convince people to accept a more open-ended view of their language capacities and repertoires.
According to Baldauf and Kaplan (2006:8) language policy and planning occur in an environment circumscribed by language ideologies arising from "historical and material situations; from socio-political and historical frameworks of power caused by discrimination and nation building". They posit that language in education is one of the major areas where the language policy of a country is seen. Like Brumfit (2006), Baldauf and Kaplan think that there is a discrepancy between the language policies put in place and actual practices in reality. Instead they contend that most policies are "diametrically opposed to reality" (2006: 9), driven by political rather than linguistic forces. This is in line with Kamwangamalu (2000) who says that there is a mismatch between language policy and language practice in South Africa, where official policy promotes multilingualism, but practices are shifting towards English monolingualism in all higher domains of use. The next section provides more evidence of the political and ideological nature of language policies in South Africa.

METHODOLOGY
Methodologically, this study forms part of Abongdia's (2013) PhD and required an analysis of the linguistic and political background of South Africa and how it relates to the establishment of the University of the Western Cape and its language policy. Furthermore, the language policy of the University of the Western Cape had to be analysed within the context of the dominant ideologies in the country as well as the ideologies of those who drafted the language policy. Finally, a cross-section of the populations (approximately 5% of the total population) at the university had to be sampled qualitatively in order to determine the current ideological positions with regards to how effectively the poli-cy was being implemented, and the impact it has on learning.

Data Collection
The data for this study was collected through the use of open-ended questionnaires completed by students and lecturers, as well as individual interviews conducted with purposefully selected lecturers, students and major stakeholders such as deans of faculties and administrators. In addition, a substantial amount of time was spent observing the research site and recording of these observations in daily field notes.
The study was conducted with the third year BA students and lecturers at the University of the Western Cape, stakeholders and administrators. The study made use of 80 participants: 12 lecturers, 5 stakeholders, 60 students and 3 administrators. The students consisted of 34 coloureds (people of mixed ancestry, predominantly bilingual in English and Afrikaans), and 26 blacks (22 isiXhosa first language speakers, two isiZulu and two Setswana first language speakers). A further breakdown of actual home language (HL) use among the coloureds students revealed that 26 indicated Afrikaans as their HL, 4 indicated English as HL while a further 4 indicated both English and Afrikaans as their HL. Naturally the research population also spoke different varieties of these languages, particularly urban, blended varieties.

Data Analysis
The qualitative data consisted of: An analysis of the language policies of the university; Written responses to a set of open-ended questions which formed part of the main questionnaire; Transcribed individual and focus group interviews and Field notes which contained observations as well as some participant responses and group discussions. These four sources of qualitative data was analysed by using Content Analysis (to uncover the different themes in the data), and Critical Discourse Analysis (to unpack the relationships between the various forms of discourses and both abstract and concrete structures of power).
In addition, careful note was taken of the respondents' tone of voice in the interviews and the manner in which the information was transmitted: passively, happily, angrily or passionately. To ensure that the emotions were well captured, the researcher personally transcribed the interviews, in order to capture the facial expressions, movements and body language of her respondents.

FINDINGS AND DISCUSSION
In analysing the language policy of UWC, note was taken of whose interests are served by these policies as well as what these policies seek to promote and or deny within the social, political and economic structures of both countries.
The South African Constitution (1996 No. 108.6) includes the following two key stipulations as part of the new democratic dispensation in South Africa, with its focus on national reconciliation and redress: (1) The official languages of the Republic are Sepedi, Sesotho, Setswana, siSwati, Tshivenda, Xitsonga, Afrikaans, English, isiNdebele, isiXhosa and isiZulu.
(2) Recognising the historically diminished use and status of the indigenous languages of our people, the state must take practical and positive measures to elevate the status and advance the use of these languages.
The language policy of the University of the Western Cape (2003) is contained in a document formally approved by the University Council, and clearly attempts to capture the spirit of these two stipulations in the Constitution. This can be seen in its aims of "ensuring equity, social development and a respect for South Africa's multilingual heritage" (Preamble of the UWC Language Policy 2003). However, English and Afrikaans are the dominant home languages of the student population, as the following breakdown of the 2010 registration statistics (the year in which this research took place) reveal: English/Bilingual English and Afrikaansspeaking: 57.3% isiXhosa: 24.7% Afrikaans: 14.9% Other SA lgs: 6.3% Foreign lgs: 4.9% This is clearly a factor that may diminish any urgency to do more to promote the use and development of formerly marginalised languages like isiXhosa at this university.  (Xhosa L1 student). Looking at the above extracts, it is evident that the use of English only medium in a multilingual institution affects students' academic performance both positively and negatively. This leaves the students with both a positive and negative attitude as the students depending on their sociolinguistic backgrounds. Furthermore, the MOI on paper does not tie with practice. This is because the students are not assisted in the other Western Cape languages (Afrikaans and isiXhosa) by lecturers and tutors who are competent users of these languages. As a result, a range of language ideologies are revealed because of the monolingual MOI.

The Hierarchy of Languages Ideology
This has to do with the belief that some languages have a higher status than others (Weber and Horner 2012:16). English is seen to be higher in status than the indigenous languages. Milroy and Milroy's (1999) notion of onenation-one language is still present, according to Anchimbe (2007), because most African countries followed the European blueprint in the design of their language policies.

The One-Nation One-Language Ideology
As research by Banda (2000), Dyers (2000), and Abongdia (2013) and others have shown that many of the students at UWC enter the university with an inadequate background in English, especially those who come from schools in poorly resourced areas. Furthermore, one is reminded of Brumfit's assertion (2006) that policy must be made known to the people concerned if one wants the policy to succeed. As Hill (2010: 42) asserts, there is a clear gap between the "stated objectives of formal language policy and actual language practice". This study is not in any way justifying the lack of policy implementation at UWC but think that dominant ideologies here may cause even the best implementation plans to fail. This therefore resulted to an evidence of how easy it was to exclude students who, as Banda (2009) argues, were never given the opportunity to learn English properly seeing that experienced and trained teachers were removed from classes after the country became a democracy.
As observed by Neethling (2010), universities appear to have tacitly ignored the SA government's 2002 recommendation on all HEIs promoting and facilitating the development of multilingualism through the use of indigenous languages. This seems particularly apt here regarding the lecturers' ignorance of the UWC language policy: "It seems to be a classic case of noble intentions of accepted recommendations simply not filtering down to the level where active steps should be taken towards eventual implementation" (Neethling 2010:67). Underlying this very basic requirement of at least familiarizing lecturers with the policy is undoubtedly the powerful ideology that English is the only suitable language for use in the academic environment of UWC. More still, the dominant na-ture of English over the country highlighted by the lecturers showed that the multilingual language policy of the country like that of UWC was not practised, leading to a mismatch between policy and practices (Beukes 2008;Brumfit 2006;Baldauf and Kaplan 2006;Kamwangamalu 2000;Spolsky and Shohamy 2000).
Taking the above evidence into consideration, one may say that the language policy of the University of the Western Cape was typical of Dor's (2004) notion of 'negotiated multilingualism'. According to Dor, the discussion on linguistic human rights (LHR) gave birth to a unique form of multilingualism called Negotiated Multilingualism. This form of multilingualism stems from the competition between global and local forces concerning the widespread use of English as a lingua franca. A position was negotiated by the policy writers between English and other indigenous languages at UWC in particular and SA in general which allowed for its use as lingua franca (Abongdia 2013). In the researcher's view, the writers of the language policy only vaguely attempted to negotiate the use of Afrikaans and isiXhosa. In this regard, the language policy of UWC conforms to a paradigm that views multilingualism partly as a human right and partly as a resource (Ruiz 1999;Abongdia 2013), thus affecting the academic performance of the students.

CONCLUSION
Given the multilingual and multicultural nature of the university, lectures are delivered in English, the exception being the different language departments. Lecturers of other departments who are competent users of Afrikaans and Xhosa are "encouraged to use the language in addition to English if it is going to ease understanding" however this is not happening. It should be noted that the policy contains a number of 'escape clauses' like 'where practicable'; 'if desired' and 'if it is practicable to do so', that could be used to justify a lack of direct action to implement the policy.
Taking the historical, socio-political and demographic factors of UWC into account, it was inevitable that the writers of this policy might have been subjected to a number of conflicting ideologies. One of these ideologies is the 'common sense assumption' that English should be the main medium of instruction and communication on this campus. At the same time, the policy also appeared to be informed by a number of ideological assumptions about multilingualism in South Africa. Although the UWC policy was written by people holding their own ideologies about the advancement of multilingualism in South Africa, the policy eventually privileged English, with two other languages, Afrikaans and isiXhosa (expected) to play 'supporting' roles. Ideologically, therefore, although the policy may have striven for linguistic pluralism, in practice it promotes linguistic assimilation and there is a very clear hierarchy of languages at this institution.

RECOMMENDATIONS
As South Africa has the highest number of official languages in Africa as well as the world, more work needs to be done to ensure that the indigenous languages are drawn upon at all levels.
At the same time, there should be much greater awareness of the ways in which most South Africans, especially the youth, make use of trans-languaging and draw on all linguistic resources in their environments in order to achieve their communicative purposes. In other words, these languages should not be treated as separate entities which are very distinct from one another.
The UWC language policy should be made known to everyone in the institution and should be accompanied by a practical and clearly-worded implementation policy, which should be open to regular reviews depending on changing circumstances. It should also be managed by an effective centre which can host regular discussions and debates on language issues. The current Centre for Multilingualism and Diversities Research might be an effective home for the policy. '