Multilingualism: Can Policy Learn from Practice

Nordic Journal of African Studies 23(1): 1–15 (2014) Multilingualism: Can Policy Learn from Practice? Rubby DHUNPATH University of KwaZulu-Natal, South Africa Michael JOSEPH University of Limpopo, South Africa ABSTRACT South African schools continue to face challenges in implementing the official bilingual policy at the level of policy development, teacher training, materials development, language pedagogy and assessment. In general, there is a growing sense of policy failure and a resignation that English will inevitably maintain its hegemony over African languages. In this paper, the authors use data generated from a Human Sciences Research Council (HSRC) research project, to document the implementation challenges schools face in enacting their language policies, highlighting how innovative schools sought to bring practice closer to policy. The paper presents a critical analysis of the work of three NGO-initiated projects that drove multilingual innovations in selected schools. These exemplary practices serve to reconceptualise the policy-practice nexus and offer an alternative to the prevailing policydriven and problem-identification research. At least one of the NGO interventions surveyed, the Home Language Project, may be considered a methodological breakthrough, with the potential to promote home languages in a low-cost, low-risk project. Keywords: multilingual education, language policy development, language hegemony, models of multilingualism. 1. INTRODUCTION The debate on multilingual education in South Africa has been re-ignited as a consequence of the absence of meaningful progress in implanting the country’s language policy, inspiring the authors to revisit data derived from a study (Dhunpath & Joseph, 2004) conducted at the Human Sciences Research Council (HSRC) in the unit for Language and Literacies Studies (LLS) in 2004. We became acutely aware of the tone of resignation around the inevitability of the failure of South Africa’s progressive and ambitious language policy, which remained trapped at the level of political rhetoric and policy symbolism. It became increasingly clear that the Government of the day had minimal interest in sponsoring a multilingual education. We therefore made a conscious decision to shift our gaze to schooling sites that had the potential to demonstrate alternative and perhaps exemplary models of practice in implementing multilingual education. Even though the research was done several years ago, we believe that many of the issues discussed and the findings are still relevant to Nordic Journal of African Studies the debates currently going on in South Africa on the implementation of Department of Education’s “Language in Education Policy” (LiEP, 1997) and such debates also continue to rage worldwide. (Alexander & Bloch, 2003; Granville et a.,l 1998; Heugh, 2003; Hornberger, 2010; Kamwangamalu, 2010; McGroarty, 2010); Phillipson & Skuttnab-Kangas, 2012. The central question that drove the HSRC research was: ‘What factors promote or inhibit multilingualism in South African schools?’ Other related questions were: • What does the official LiEP expect of schools? • What multilingual practices do schools claim to implement? • How consultative are schools in seeking parents’ language preferences for Medium of Instruction (MoI) and Additional languages (ALs)? • What processes did schools follow in formulating their language policy? • What forms of support are given by agencies such as the Department of Education (DoE), Higher Education Institutions (HEIs) and NonGovernment organizations (NGOs)? • To what extend do curricular and classroom practices in content classes affirm the use of African home languages for the learning of content? • What have schools done to foster multilingual learning environments? It would be impossible to present findings related to all of these questions in the space of this paper. Important insights will therefore be selectively discussed with the hope that the breadth and depth of the research will be reflected in the analysis of the themes that emerged and the generalizations offered. The key question explored in this paper is: What are the implementation challenges schools experience in enacting their language policies and how do innovative schools bring practice closer to policy with reference to the work of selected Non-Governmental Organisations (NGOs) that drove multilingual innovations in selected schools? 2. RESEARCH DESIGN: SITES, DATA-COLLECTION, INSTRUMENTS AND DATA ANALYSIS This section spells out aspects of the research design, broadly conceptualized within a mixed methods framework but leaning towards a qualitative approach, which enabled narrative and reflective accounts, flexibility in data collection procedures and data analysis that enabled the generation of new research insights. The schooling sites were purposively sampled to focus on those that were, as far as possible, linguistically heterogeneous, have significant numbers of learners with an African language as mother tongue, represent average lower middle-class or working class communities and receive some sort of support in policy development and implementation. The sampled were limited to those that had received support from the following three service providers: 2 Multilingualism: Can Policy Learn from Practice? 1. The English language Teaching Centre (ELTIC) - a well-known language development NGO located in Gauteng, which ceased to function several years ago as a result of curtailed funding. Notwithstanding its demise, the authors have included this initiative in this paper because it is regarded as one of the most successful models of promoting multilingualism worthy of being re-activated. 2. The Home Language Project (HLP) - a new, relatively unknown but functioning service provider in Gauteng. 3. The English Language Education Trust (ELET) an NGO that has been involved in language development for more than 20 years in KwazuluNatal. A total of 19 schools (mostly urban and peri-urban), 14 in Gauteng and 5 in Kwazulu-Natal were selected. They ranged from urban to rural and farm schools, from primary to high schools, and from all male to mixed gender schools and all had received support from NGOs in policy development and implementation of multilingual practices through language teacher development, curriculum development and materials development. Data-collection instruments included (i) a school profile questionnaire, completed by school principals, to obtain information on the demography of the school, languages of teachers and learners, languages offered at the school, the availability of multilingual resources and the participation of parents and learners in the development of the school’s language policy (ii) interviews with educators to establish the levels of proficiency and competence in the MoI, perceptions of and attitudes to multilingual teaching/learning, and to gain teachers’ understanding of the process of policy development in the school (iii)focus group interviews to gain insights into learners’ attitudes, perceptions and preferences, focus group techniques were used, ensuring representativeness in terms of ethnicity, gender and grades. Informed consent was sought of all respondents (iv) interviews with School Governing Body (SGB members) to ascertain their attitudes to multilingual education and their perceptions of school language policy development and implementation; (v) classroom observations mainly of content classes, which ranged from the foundation phase to secondary schools, but with the focus predominantly on the intermediate phase. The reason for choosing content classes for observation was they represent the highest level of multilingual education (use of the home language to access academic knowledge). The enquiry into content (subject) classes was mainly to examine the highest level of multilingualism, namely the issue of medium of instruction. Additional data were drawn from field notes, documents and transcripts of selections from tape-recordings and photographs. However, due to logistical issues, no classroom observation was done in the 5 ELET schools in Kwazulu Natal and document analysis of the language policies (where they exist) of the selected schools and related documents to establish the process through which the policy was developed. This included an analysis of the levels of participation by 3 Nordic Journal of African Studies different sectors in the school community to gauge the level of implementation of multilingual education as envisaged in the LiEP. Eight researchers1, including the two authors, conducted the fieldwork over a period of twenty days in the 19 selected schools. A multi-stage data analysis was employed to develop thick descriptions of the schooling sites and to document the attitudes, perceptions and experiences of educators, learners and parents. Case studies of each school were evolved from the primary data collected and analysed to yield generalizations based on four broad themes. 3. MAJOR FINDINGS The four themes are interrelated and overlap in part but together they help to recover the route we have tracked from policy development to practice. 3.1 THEME 1: POLICY DEVELOPMENT PROCESSES In answer to the research question, ‘What processes did schools follow in formulating their language policies?” The data revealed that processes ranged from: decisions taken a priori without any consultation, to more or less democratic forms of policy development. In some cases School Governing Bodies (SGBs) were excluded altogether and were only brought in to legitimize an already formulated document. In other cases, SGBs played a leading role in collecting parents’ views on language choice. The most autocratic models of policy formulation were those in which a priori decisions were taken about English as MoI, and learners who may have had other choices were excluded from exercising any choice. Learners were also forced to sign contracts that ensured that the language policy could not be changed at a later date. These policy processes showed a lack of responsiveness to the community and is in direct opposition to official policy. Though only a few schools openly admitted to an a priori selection of English as MoI, other schools implicitly signalled their choice of English and the SGBs subsequently endorsed it. In other words, except for a very few schools, choice was absent. It is important to note that the absence of choice only relates to MoI. However more democratic processes were followed in relation to additional languages, but these were based on the highest votes for a majority African language. The home languages of smaller language groups were therefore excluded. Without exception all schools, excluded learners’ views from surveys to establish language preferences. Our focus group interviews with learners in the 19 schools reveal that learners do have clear, honest and very specific views, 1 The six other HSRC researchers are Krish Govender, Matthews Makgamatha, Makola Purutsee, Brutus Malada, Margaret Omidire and Heidi Paterson. 4 Multilingualism: Can Policy Learn from Practice? and articulate them well. They were unambiguous about their right to learn their own languages but had more complex views about their languages as resources for learning. Learners’ views were found to be very divergent from those of adult respondents (principals, SGBs, educators and parents) and their concerns about ‘passing’ Matric were very real. A range of communication procedures, were utilised including providing parents drafts of proposals for approval, to using questionnaires or using wordof-mouth procedures. The exclusion of the views of non-literate parents (either in the home language or English) is a serious problem that needs to be addressed. Multilingual practices outside the classroom for informal purposes enjoyed the support of all role players in schools’ policy formulation. There was however no evidence, except for one school, of schools seeking parents’ or educators’ views on creating a multilingual print environment. No alternative provisions were sought from parents and educators for the provision of home language learning in the former ELTIC schools but many of the HLP schools made provision through one class per week per school. These HLP classes catered for all the home languages of learners from all grades in all seven of their schools. The HLP has also started biliteracy extra-mural programmes for home language and English reading for pleasure in some junior primary schools. The HLP is now one of the most successful models of bilingual education at school level. 3.2 THEME 2: SUPPORT FOR POLICY The ‘delivery’ of policy seems to be shared by schools and Education department district officials in the following way: schools are expected to formulate their school policy in democratic ways, implement their policy (as MoI, AL and creation of a multilingual environment themselves) with the Education department offering to facilitate through feedback on policies and training inputs. However, our study showed that in actual practice the majority of the schools surveyed were left to their own devices, with school principals and subject heads giving leadership. Principals with a clear vision and commitment proved decisive in whether schools developed a progressive language policy. Interventions from service providers also depended crucially on how open school principals were to such interventions. Most schools were unhappy about the training workshops given by district officials and complained that the training concentrated specific to learning areas. Multilingual approaches and Language Across the Curriculum perspectives were missing from these workshops. Moreover, workshops were seen by some educators to be too abstract and not particularized for classroom practice. One principal’s view appears constructive and worth reporting at some length. She argued for a two-level training workshop: an initial language-across-the5 Nordic Journal of African Studies curriculum training for staff of all learning areas, within which the role of multilingualism would be addressed and then, learning area workshops infused with multilingual possibilities for learning of content integrated with English. Such workshops, the principal suggested, would help to overcome a fragmented approach to curriculum development and enable staff to develop a common vision of the relation between policy and their own curriculum. There is a link between such workshops and the focus of the current research on exemplary practices. The inclusion of exemplary multilingual practices in training workshops would enable a shift from policy-driven to exemplar-driven multilingual training. 3.3 THEME 3: ATTITUDES OF RESPONDENTS This section consolidates various findings related to the perceptions of different stakeholders to the use of English and African languages as MoIs/ALs. Attitudes to English as MoI and African languages as MoI showed the highest divergence among various respondents. Overall, English was supported for instrumental/rational purposes and African languages for social solidarity. In terms of Ruiz’ distinctions (1984) this suggests strong support for African languages as a right but not as a resource. In relation to educational domains, English dominates the instrumental domains, namely as MoI for content subjects whereas African languages are supported in the domains of ALs and on the playgrounds and in staff rooms. When respondents’ attitudes were sought about the use of African languages in the instrumental domains (for teaching and learning content subjects) African languages were seen as a problem. We concluded that the distribution of English and African languages across domains, with English in the ‘higher’ instrumental domains was how respondents ‘equalized’ the two languages, maintaining English as a resource and African languages as a right. However, in response to a very specific set of questions presented to five learners at a focus group discussion, interesting preferences emerged. Rather than ask the binary question: English or African language as MoI, the learners in this group were presented with three models: I. II. III. English only for all subjects Dual-medium education Sepedi only for all subjects. All learners chose option II: Dual-medium education. They were then asked if they preferred (i) (ii) (iii) a 50/50 distribution of English/Sepedi as MoI 80/20 of English/Sepedi or 20/80 of English/ Sepedi. 6 Multilingualism: Can Policy Learn from Practice? Four learners chose option (ii), and one learner chose option (iii). This smallscale initiative seems to suggest that research needs to be based on real possibilities and micro-level dynamics rather than policy recommendations. Additional languages as subjects were celebrated as enabling a choice between Afrikaans and African languages. This did not however solve the problems of all African-language-speaking learners. Those coming from urban primary schools (into high schools) where only Afrikaans is offered continue to take Afrikaans, unhappily, but assuming that it was safer to continue with a language they had already invested in. The major divergence here was about the choice of a majority African language as AL versus other home languages. Many learners and parents of minority language groups were unhappy with the choice of a majority African language as AL and felt their own home languages were being undermined. In principle, school authorities were willing to offer home language as additional languages if they had adequate numbers of students and governmental support. Only the Home Language Project has managed to provide for home languages in a low-cost, low-risk project. Both urban and rural schools enthusiastically supported the idea of reading for pleasure. However, school libraries were often not functional, nor wellresourced in African language books. Learners and parents wanted more opportunities to read. Sadly, African language-speaking parents had no more than ten books in their home language as compared to the much higher number of English books recorded in the household. Both parents and learners wanted more books in African languages in school libraries and more access for learners. 3.4 THEME 4: CLASSROOM INTERACTIONS Data from the classrooms revealed the following: Rural schools used African languages as resource for learning within a de jure English MoI situation. Evidence of code-switching and code-mixing for social functions and for processing academic concepts was found in abundance in rural schools, and to a lesser extent, and in more teacher-regulated forms in urban schools. There was greater conformity to the demands of Outcomes Based Education (OBE) than to Multilingualism. Classroom print environments were almost entirely English only. Some classes showed multiple home languages being used in a complex set of interactions (pair work, group work, learner-whole class interaction, teacher-whole class interaction) on tasks of varying degrees of cognitive challenge around academic reasoning. We saw these practices as exemplary as they show African languages being used to promote higher-order thinking. However, these practices are not valued by the teachers themselves or even by the learners as a form of progressive pedagogy. The insight we derived 7 Nordic Journal of African Studies is that when learners and teachers focus on the content of learning in order to grapple with meaning, they are unconscious of the languages that are used. In such lessons, where the comprehension and production of academic content is the focus, the language of teaching is actually the language of learning, in the form of code mixing and code switching. Observation of African languages as subject (ALs) in the HLP classes showed marked innovations in the teaching of African languages from conventional lessons. The most striking feature of these lessons is the use of the home languages of all the learners. The home languages focus on contemporary genres (such as news reports) and use both analytical and dramatic modes focused on literacy materials (that are either in English or in African languages). High levels of peer interaction and teacher-learner interaction occurred in these lessons. High levels of cognitive reasoning were also observed. The HLP innovations in additional language teaching constitute not only a methodological breakthrough but also a breakthrough in terms of visibility, through videos and research. 3.5 DATA ARISING FROM THE SPONTANEOUS EXPRESSION OF RESPONDENTS’ FRUSTRATIONS The findings reported in section 3 emerged from the research design. However, respondents felt the need to share their frustrations. This served as a further source of data. Several of their frustrations were related to logistics: school timetabling issues, excessive assessment expected in the curricula, lack of governmental support in training workshops, especially the neglect of training for multilingual education, etc. References to these practical problems often turned out to be more revealing than responses to the planned questions. Using these logistical problems as an entry point, our probing suggested that respondents’ reports often appeared to be self-contradictory. Rather than dismiss their views as inconsistent, a deeper analysis showed an honest conflict within their minds about multilingual education. We examine below these conflicts at a more general level in terms of the paradigms of the politics of language and of language/s in learning. The paradigms theorize some aspects of the data and the frustrations expressed. 8 Multilingualism: Can Policy Learn from Practice? 4. PARADIGMS THAT UNDERLIE POLICY AND PRACTICE 4.1 THE CONFLATION OF THE ‘HEGEMONY OF ENGLISH’ AND ‘MONOLINGUALISM’ Many parents (both black and white), teachers and principals expressed views that favoured an ‘English only’ approach and saw English as the only suitable language for expressing modern knowledge, along with the view that African languages are unsuitable for these purposes. This view is in line with the more general patterns reported on by Heugh (2002). We interpreted these views as support for the hegemony of English. However, these same respondents were quite aware and disturbed about the widespread displacement of African languages by English. This was especially the case with the black respondents, and even more so with the few progressive black parents who saw multilingualism as a resource for their children, drawing empathetically upon their own identities as African language speakers. White parents likewise were aware of the displacement of African languages and expressed regret that they did not know an African language. Often these self-contradictory views were not overtly stated and were also not visible as contradictions to the researchers. However, in one telling interview, a black parent told one of the researchers (Joseph) that African languages must be used more in the school, but almost in the same breath added that English should be used by the children ‘in the playground also’. This highlighted a concealed contradiction around the status of English and African languages in the education domain. These conflictive views at least confirmed that respondents were not trying to give politically correct answers. Our interpretation of why they presented a self-contradictory view on English vis-à-vis African languages was that they were conflating two different (though related) concepts: the hegemony of English, and monolingualism. Respondents were clearer about the hegemony of English and wanted at the very most, English to have a dominant (but not hegemonic) status. But they could not entertain the idea of two languages as MoIs. In their view, instruction could only be monolingual. With such a covert endorsement of monolingualism for instructive purposes, English and African languages confronted them as a choice between the two. English appeared more favourable to them. The respondents’ confusion becomes clearer when set against theoretical knowledge of how languages acquired power in history. From a theoretical (rather than folk) perspective, multilingual scholars distinguish on the one hand between the power of languages in relation to each other, and between monolingualism and multilingualism. Though the two concepts are different, the history of colonialism saw the rise to power of European languages on the back of monolingual national policies. Many of the European languages had to achieve internal conquest over other national languages and dialects (through standardization) and the myth of one language as symbolic of national unity 9 Nordic Journal of African Studies served this purpose. The power that colonial languages had over the nonEuropean languages of Africa, Asia and Latin America also conveyed the idea of monolingualism as an advantage for these subject countries. It was convenient to ignore the fact that not all European countries were monolingual (e.g. Belgium, Switzerland etc.). Languages such as English were not only seen as more powerful and more desirable (by non-English subjects) but monolingualism was seen as synonymous with English education. But monolingualism, as we well know, undermines all forms of diversity, and is therefore more subversive. In the South African context, the opposition to the hegemony of English is not matched with an equal opposition to monolingualism. The former is more challenged than the latter. This suggests why respondents can oppose English at the political level but allow it to re-enter at the educational level via monolingualism. 4.2 A DICHOTOMOUS VIEW VS. AN INTEGRATED VIEW OF LANGUAGE CHOICE A further explanation for the confusion in respondents’ minds come from what may be termed ‘the dichotomous view’ of language choice. Heugh (2002:180), critiquing the claims of supporters of monolingualism, points out that they make a false dichotomy between English and African languages, by presenting them as languages to choose between. Such a view has been the basis of national research (such as Taylor & Vingevold, 1999) and has yielded the misperception that parents went straight for English. When, on the other hand, as Heugh argues, research is conducted on the basis of an ‘additive’ view of multilingualism, or what we would call an ‘integrated’ view, the choices that parents make is more complex, and shows that they reject a straight-for –English in favour of a more gradual transfer to English. Unfortunately because of the prevalence of a language rights laissez-faire approach to multilingualism in the post-apartheid South Africa, the monolingual myth that parents have opted for a straight-for-English approach retains its strength through the dichotomous view of language choice, and has percolated the minds of African-language speaking parents as common sense. Advocacy and research must therefore challenge the dichotomous view by avoiding the ‘either/or’ monolingual axiom and replacing it with questions that show multilingual inclusivity, through the integrated approach. Hypothetical models of dual-medium instruction were presented to a small focus group of learners reported in section 3.3. Learners’ responses revealed that they favoured a bilingual model, even if it was one of unequal uses of the two languages. As Heugh points out, “it is not a question of an either English or African languages debate. It is about ensuring that pupils access fairly and equitably the content of the curriculum.” (2003:19). The issue of learning two 10 Multilingualism: Can Policy Learn from Practice? languages very well is only secondary, compared to the successful acquisition of content. In other words, when the learners’ attention is drawn to the issue of content, language choice is seen as integrated, rather than dichotomous. This is because learners are now viewing languages as resources for knowledge attainment, and not the status attached to each language. Some implications for language advocacy and research emerge from the interpretations offered above. An advocacy that plays down the politics of the hegemony of English and instead offers educational arguments about the advantages of a bilingual education is likely to resolve the false dilemma of languages as dichotomous choices. Research methodology likewise needs to frame questions that will distance themselves from political rhetoric and false dichotomies and draw upon integrated additive models of language in education. Such theoretical models backed by practice is to be found in the work reported by Alexander & Bloch (2003), Ramani & Joseph (2008), Hornberger (2003), Hornberger & McKay (2010). 4.3 CODE-SWITCHING FOR PHATIC OR COGNITIVE PURPOSES? The de facto use of code-switching (use of two languages in the same classroom) was observed and recorded, and attitudes of teachers across schools were also investigated. The results showed variation in attitudes and actual use of code-switching. In the linguistically homogenous schools (comprised of only black students) there was more (as can be expected) of code switching both quantitatively and qualitatively, in contrast with heterogeneous schools, where there were both black and white students. The generalization we make from both kinds of schools is that respondents (teachers specifically) think of language policy only at the level of the medium of instruction (MOI) and not the languages of the learning and teaching (LOLT). Their attitudes to code-switching as well as practices of code-switching in their classrooms hide a covert language in education policy. African language-speaking teachers in the rural schools use and ‘allow’ their students to use code-switching almost as the norm of classroom interactions. They do not however attach importance to the de facto use of the home language as a resource for learning, and do not see it as part of official LiEP. On the other hand, some of the teachers in the heterogeneous urban schools (from the HLP sample, but untrained by HLP) claimed to tolerate code-switching as a school language policy, but when it came to their own classrooms, white monolingual teachers offered various reasons for forbidding its use, such as the possible abuse of the home language for non-learning purposes, and fears that it might affect group solidarity in mixed ethnic groups. The use of peer translations as a solution was not entertained by these teachers. However, HLP-trained teachers in the HLP School gave full value to code-switching and ensured its practice. 11 Nordic Journal of African Studies Classroom observations confirmed that code-switching was used for social (phatic) as well as cognitive purposes. But teachers did not appear to notice this distinction. In some of the heterogeneous schools, for instance, the social use only of code-switching was commended as upholding multilingualism. Our conclusion is that the official LiEP does not empower (or pressurize) teachers to exercise choice about LOLT sufficiently, and places a much greater emphasis on the MOI. This is unfortunate, because it is LOLT that is within the power of teachers. An official language policy needs to empower teachers at the level at which they naturally have power, so that they could transform their de facto multilingual practices into de jure official policy. It is only in this way that official policy can convert and complement the implicit policies underlying teacher practices, and avoid appearing as the alienating bureaucratic discourses that most teachers perceive them to be. Of importance too is the need for a cognitively-oriented approach to codeswitching that views it as part of what Cummins (1996) calls the common underlying principle. If code-switching is used for reasoning through content, then such reasoning transfers into English as well and does not have to be relearnt in English. Cummins’ model (1996) of the four quadrants of language proficiency is thus of great value in distinguishing between these two important functions of code-switching. 5. MODELS OF MULTILINGUALISM In conclusion, we examine the approach to multilingualism of the various schools that we researched in terms of models advanced in South Africa (see Heugh 2003). Though we are aware of the many models of multilingualism that proliferate the literature, this paper is concerned with locating the nexus of language policy and practice in the schools we researched in terms of the cline of ‘assimilationism’/ ‘additive multilingualism’. Assimilationism refers to the complete surrender to English as the medium of instruction as well as other domains, whereas additive multilingualism aims at gaining access to a dominant language (English in South Africa) while preserving and developing the mother tongue for academic and content acquisition. The typology of multilingualism we derived is presented in table below: A. Assimilationist B. Transitional D. Strong Additive Multilingualism C. Towards Additive Multilingualism 12 Multilingualism: Can Policy Learn from Practice? In the light of these models, it is clear that no school in the HSRC research approximated to quadrant D, which is the LiEP recommendation. For African language-speaking children all South African classrooms from the post grade 3 level could be regarded as ‘assimilationist’ on the grounds that the official MOI is English, and therefore displaces the mother tongue from accessing school content at the highest level, namely assessment. This also applies to the samples of schools chosen for the HSRC research. Though there is value in upholding quadrant D as an ideal, it is counter-productive to present models of multilingualism as polarized into assimilationist and additive, as such polarities would ignore the more intermediary forms of multilingual teaching and learning that take place in the many different domains of school life. To ignore these domains would not only over-simplify the case for multilingualism, but also undermine the potential for teachers to perceive themselves as agents of multilingual educational change. Many of the heterogeneous schools we observed were characterized by assimilationist practices (Quadrant A) though they professed a wish to be multilingual in some form or the other. Significantly, several schools refused to participate in our research. These included some Afrikaans-only schools. Assimilationism of a different form may exist for these schools, namely assimilationism of black students to Afrikaans mainly and English additionally. Some of the white urban schools practiced a subversive form of assimilationism by ensuring an English only policy at the admission stage itself. Students (parents in fact) were thus effectively silenced from articulating their multilingual rights (if they had any) from gratitude that they had obtained admission. These schools did not affirm multilingualism either at the level of policy or practice even as a wish. Most of the urban schools from our sample were not of this extreme form of assimilationism, and offered at least symbolic support for multilingualism. The more progressive of the urban schools did not actively encourage the use of African languages but tolerated it as a subject and allowed its use on the playground. They saw this as a form of multilingualism. From our additive multilingual perspective, this appears to be a form of transitional multilingualism to be tolerated till students master the English language. In homogenous (usually rural) schools, the home language was used because there was no alternative. The home language was tolerated but not valued as noted earlier. Thus, this de facto form of multilingualism also is transitional, as even if its use persisted, it was not valued and did not contribute to the self-image of teachers and learners, who on the whole (particularly) teachers aspired for an English education. It was only the NGO-led schools that fit into Quadrant C and offer hope for viable forms of bilingual education. 13 Nordic Journal of African Studies 6. CONCLUDING COMMENTS In terms of the Ruiz model of indigenous languages being seen either as problem, right or resource, it would seem that most of the schools in Quadrant A see African languages as a problem, those in Quadrant B see it as a right and perhaps only those in Quadrant C see it as a resource for learning. In most rural schools teachers and learners are using their own languages at least as a partial resource. This partial resource is a cognitive one that is theorized using Cummins’ framework (1996) as the use of home languages for scientific reasoning. If teachers of these schools saw themselves as innovative like the teachers of the HLP, and if government policy recognized and rewarded such schools for these practices there is no doubt that a policy from below could contribute to an evolving multilingualism to complement multilingual policies from above. REFERENCES Alexander, N & Bloch, C. 2003. A Luta Continua!: The Relevance of the Continua of Biliteracy to South African Multilingual Schools. In: Hornberger, N. (ed.), Continua of Biliteracy: An Ecological Framework for Educational Policy, Research, and Practice in Multilingual Settings. Cleveland: Multilingual Matters. Cummins, J. 1996. Negotiating Identities: Education for Empowerment in a Diverse Society. Ontario, CA: California Association for Bilingual Education. Dhunpath, R & Joseph, M. 2004. Factors promoting or inhibiting Multilingualism in Post-apartheid South African schools. Unpublished Research Report by the unit for Language and Literacies Studies: Human Sciences Research Council, Pretoria, South Africa. Granville, .S, Janks, H., Joseph, M., Mphahlele, M., Ramani, E., Reed, Y. & Watson, P. 1998. ‘English with or without g (u) ilt’. Language and Education 12(4): 254–72. Heugh, K. 2002. The case against bilingual and multi-lingual education in South Africa: laying bare the myths. Perspectives in Education 20(1), 2002. 2003 Language Policy and Democracy in South Africa. Stockholm University: Centre for Research on Bilingualism. 14 Multilingualism: Can Policy Learn from Practice? Hornberger, N. 2003. Continua of Biliteracy: An Ecological Framework for Educational Policy, Research, and Practice in Multilingual Settings. Cleveland: Multilingual Matters. Hornberger, N. 2010. Can Schools Save Indigenous Languages? Policy and Practice on Four Continents. Palgrave Macmillan: 175 Fifth Avenue, New York, NY 10010. Hornberger, N. & McKay, S.L. 2010. Sociolinguistics and language education. Clevedon, UK: Multilingual Matters. Kamwangamalu N.M. 2010. Multilingualism and codeswitching in Education. In: Hornberger, N. & McKay, S. (eds.), Sociolinguistics and language education, pp. 116–142. Clevedon, UK: Multilingual Matters. McGroarty M.E. 2010. Language and Ideologies In: Hornberger, N. & McKay, S. (eds.), Sociolinguistics and language education, pp. 3–39. Clevedon, UK: Multilingual Matters. Phillipson, R. & Skutnabb-Kangas, T. 2012. Linguistic and Cultural Genocide and Crimes against Humanity in Education. Paper presented at the UTLO Seminar Series, 10 October, 2012, University of Kwa Zulu Natal, South Africa. Ruiz, R. 1984. Orientations in language planning, NABE journal 8(2): 15–34. Ramani, E. & Joseph. M. 2008. Achieving academic excellence through higher-order thinking in an African language. Perspectives in Education 26(4): 42–56. Taylor, N. & Vingevold, P. (eds.) 1999. Getting learning right: Report of the President’s Education Initiative Research Project. Johannesburg: the Joint Education Trust. About the authors: Rubby Dhunpath is the Director of Teaching and Learning at the University of KwaZulu- Natal South Africa. He has researched and published in education policy, language policy, life-history methodology and international doctoral education. Michael Joseph is a senior researcher, located at the University of Limpopo, South Africa, working on SANPAD and NRF multilingualism projects: His research interests include language policy, bilingual education and language curriculum development. 15