Mother tongue-based bilingual education (MTbBE) is an educational approach that uses the home language (HL) of learners as a resource concurrently with English in the classroom to enhance meaning-making while improving their academic achievement (August-Mowers 2025; Department of Basic Education [DBE] 2024). The implementation of MTbBE is supported by the South African Constitution and the Language in Education Policy (LiEP) of 1997, which clearly states that all 11 official languages should have equal status (LiEP 1997). This means that the goal of LiEP was to ensure social cohesion in schools by emphasising the importance of languages co-existing and being given equal recognition, with an emphasis on developing multilingual environments (Makupe & Machaba 2024; Sibanda & Tshehla 2025). In essence, in the South African context, MTbBE aims to promote additive bilingualism and the development of African languages in education. However, for years, that practice was only visible in the Foundation Phase (FP), where learners were only taught in their mother tongue for the first 4 years in Grade R-3, then a switch to English instruction from Grade 4 upwards (Makoe 2014; Masunungure & Modiba 2023; Sibanda & Tshehla 2025). Mother tongue instruction in the FP is deemed a suitable strategy in the CAPS, as learners need a firm understanding of their mother tongue before learning an additional language, as they can transfer the skills they have learnt in their mother tongue to the additional language (Cummins 2000). Nevertheless, the implementation of MTbBE in schools sought to close that gap through providing an extension of mother tongue instruction beyond the FP, through implementing a bilingual approach that advocates for educators to use African languages together with English as the Language of Learning and Teaching (LoLT) in schools, since African languages continue to be under-represented in public settings, including schools (Kretzer & Kaschula 2021). Premsrirat and Burarungot (2022) add that inclusion in multilingual and multicultural contexts requires equitable access to quality education, as indicated in the Sustainable Development Goals (2022), and the implementation of MTbBE carries that agenda by emphasising that African languages should be used beyond the FP as LoLT.

The MTbBE project has started across the nine South African provinces, with a target in Mathematics, Natural Science and Technology in Grade 4 and Grade 5. This is because of the officialisation of circular 18 of 2024, which stipulated that in 2025, official enrolment of MTbBE should start in schools (DBE 2024). Mother tongue-based bilingual education implementation is regarded as a transformational and progressive policy shift in the post-apartheid era, since it seeks to redress the misinterpretation of LiEP (Singh 2025). Studies on MTbBE implementation have gained prominence worldwide, in countries such as Ghana, Nigeria and Thailand, and have revealed the positive effects of this bilingual approach in fostering social cohesion (Igboanusi 2008; Premsrirat & Burarungrot 2022; Owoo 2022).

Mother tongue-based bilingual education is also supported by three United Nations Educational, Scientific and Cultural Organisation (UNESCO) principles: support for mother tongue instruction, bi/multilingual education and the use of language as an essential component of intercultural education (UNESCO 2003). The implementation of MTbBE is also a response to international studies such as Progress in International Literacy Studies (PIRLS) which indicated that South Africa (SA) learners performed poorly in international tests as compared to their counterparts, and one of the suggestions was a need to change the LoLT, and introduce one that will accommodate most learners; since research has shown that the learners’ learning challenges are language-related (DBE 2024; Masunungure & Modiba 2023).

However, although studies on MTbBE in SA have been conducted at a larger scale in both higher and basic education, and their good practices are known to the public (Mashige, Cekiso & Meyiwa 2019; Motloung, Mavuru & McNaught 2021; Singh 2025), these studies have not focused much on educators’ narratives in township schools regarding language choices and how the inclusion criteria used to implement MTbBE affect learners’ overall performance, given that linguistic diversity in schools has become a serious concern (Masunungure & Modiba 2023).

Studies on the effect of linguistic diversity on learner performance have been conducted (Masunungure & Modiba 2023), but they did not focus on the implementation of MTbBE and the inclusion criteria used in multilingual township schools to select schools’ language policy. Hence, the specific focus of this article is to present educators’ narratives of implementing MTbBE in township schools, drawing on studies conducted worldwide and nationally. It presents limited data from township pilot schools with multilingual learners. It focuses mainly on educators’ narratives about the language inclusion criteria for implementing MTbBE and how these criteria affect educators’ instructional practices and multilingual learners’ academic performance. This article seeks to bring educators’ narratives to the fore in the implementation of MTbBE in multilingual township classrooms by exploring how teachers interpret and apply inclusion criteria in their school’s language policy within MTbBE, and how these interpretations are shaped by multilingual classroom realities and broader language-in-education policy tensions in South African schools, through answering the following research questions:

This study adopted a qualitative research approach, as it is concerned with gathering deep insights from participants’ experiences and understanding complex phenomena (Ahmed 2024; Lim 2025). Qualitative research was deemed appropriate for this study, as it seeks to report educators’ insights on implementing MTbBE in multilingual township schools, with a focus on the inclusion criteria used in their schools to select languages for MTbBE implementation. Qualitative research was also chosen because of its ability to report on human experience and perspectives through rich textual data (Lim 2025). Within qualitative research, a case study was used because of its unique ability to reveal the dynamics of a phenomenon in its real-world context and align with the researcher’s objectives (Lim 2025). For instance, this study specifically focused on understanding how schools’ language policy inclusion criteria are used in multilingual township schools, and how this affects educators’ practices when implementing MTbBE.

Purposive sampling was used to select six educators from multilingual township schools who are piloting MTbBE from Grade 4 upwards. For educators to be included in this study, they had to either teach Maths or Natural Science and Technology in a multilingual township using the indigenous language, as this constitutes the piloting of MTbBE in primary schools. In essence, this study employed purposive sampling to select participants who possessed direct experience and knowledge relevant to the implementation of MTbBE in township schools. Since purposive sampling is the intentional selection of participants that offers desirable insights into the phenomenon (Tajik, Golzar & Noor 2025), the study selected six participants, as only a limited number of schools were piloting MTbBE. Township schools were purposefully chosen because their context reflects true and flexible multilingualism, and that allowed the researcher to gain a deeper understanding of how the language inclusion criteria affect educators’ practices and learners’ performance. Purposive sampling was deemed appropriate for the study, as it aimed to gain rich, context-specific insights from educators directly involved in the practices of multilingual classrooms and learner inclusion processes. Six teachers were selected based on criteria such as experience in township schools, working in multilingual classrooms, and familiarity with MTbBE practices. The qualitative nature of the study sought depth of understanding rather than statistical generalisation. The sample size was considered adequate, as purposive sampling prioritises depth of understanding rather than statistical generalisation; hence, it involves selecting a desirable population that serves the study’s purpose (Etikan, Musa & Alkassim 2016). Notably, in this study, most educators were bilingual, with more than 3 years of teaching experience in township schools situated in a multilingual context, as demonstrated in Table 1, although they were teaching learners who were multilingual, as they alluded to during their interviews.

Data were collected through semi-structured interviews with educators teaching in multilingual township schools, where linguistic diversity is the norm in the classroom (see Table 1) and choosing one language over another often excludes other learners. Using semi-structured interviews was considered the best fit, as they offer flexibility in questioning, allowing the researcher to adapt questions based on participants’ responses and, in turn, enabling participants to provide in-depth information about the phenomenon under investigation (Chand 2025). In addition to semi-structured interviews, document analysis was used as a secondary data collection method. Document analysis in qualitative research involves analysing existing texts, with a clear focus on the text under analysis (Chand 2025; Morgan 2022). For instance, in this study, the documents analysed included the workbooks educators received during the MTbBE training course and the language plan for each topic. Document analysis is about systematic examination and interpretation of documents to gain insights about a specific phenomenon (Chand 2025). This study employed methodological triangulation through semi-structured interviews and document analysis. Semi-structured interviews were conducted to gather in-depth insights into teachers’ experiences, interpretations of MTbBE implementation and perceptions of the inclusion criteria in MTbBE. Relevant policy documents, such as curriculum statements provided in MTbBE training, were also analysed to contextualise teachers’ accounts and compare these with official policy expectations. Triangulation of these two data sources increased the credibility of the findings by enabling cross-referencing between teachers’ reported practices and policy intentions. It also enabled a more comprehensive understanding of how inclusion criteria are understood and enacted in practice. Nevertheless, although interviews and document analysis enabled methodological triangulation, the study did not include classroom observations. As a result, findings primarily reflect teachers’ reported practices and policy texts, rather than directly observed classroom interactions.

Trustworthiness is crucial in qualitative research (Ahmed 2024), and to ensure trustworthiness in this study, member checking was done after transcribing data collected through semi-structured interviews. Member checking allowed participants to review and confirm whether the transcript accurately reflected their viewpoints (Ahmed 2024). Triangulation was also applied in this study, where data were collected through two datasets, semi-structured interviews and document analysis, and using these two datasets allowed the researcher to cross-verify the findings (Ahmed 2024). In instances where educators mentioned how the bilingual papers were presented, the researcher also had to request samples of those papers to verify their claims. Using two datasets also reduced potential bias arising from reporting data from a single source (Ahmed 2024). Data were analysed thematically following Braun and Clarke’s guidelines, which consist of familiarisation with the data, generating initial codes, searching for themes, reviewing themes, defining and naming themes and producing the report (Braun & Clarke 2006). Firstly, the researcher familiarised themselves with the data by conducting verbatim transcriptions of audio-recorded interviews, allowing for the identification of recurring patterns relevant to the research questions. Secondly, the researcher generated initial codes from the data deductively, guided by the study’s research questions and the study’s theoretical framework of translanguaging, while remaining open to patterns emerging from the data. Thereafter, the researcher began searching for themes by grouping relevant data into each theme, leading to the naming of themes based on how they responded to the research question. The last step was to produce the report by relating themes to the research questions and providing quotations where necessary. Following these guidelines allowed the researcher to organise the themes strategically and in accordance with the research questions. Reflexive practice was employed throughout the analysis to critically examine how the researcher’s own assumptions and interpretations might impact the analytical process. Ethical standards were upheld, and pseudonyms were used instead of participants’ real names (see Table 1).

This section presents the findings of this study under three sub-themes that emerged from semi-structured interviews with educators and document analysis. The identified themes were the effects of language inclusion criteria in a multilingual township school; positive aspects identified by educators in implementing MTbBE; and challenges posed by the inclusion criteria in implementing MTbBE; these themes correspond with the research questions, which focused on educators’ narratives on the inclusion criteria used for MTbBE in multilingual township schools. Direct quotes were used to present their views and experiences.

This study used a small-scale qualitative inquiry, since it aimed to explore educators’ experiences in a context-specific area rather than produce broadly generalisable conclusions. The translanguaging lens was employed to examine how the inclusion criteria for selecting languages for MTbBE implementation in multilingual schools shape educators’ instructional practices. The study was conducted in township primary schools characterised by significant linguistic diversity. The schools served communities where multilingualism formed part of everyday social life. Despite the multilingual nature of the communities and the implementation of MTbBE to elevate the standard of indigenous languages, English continued to be associated with academic progression and socioeconomic mobility, influencing language practices and perceptions within the schools. This study revealed that while opportunities for translanguaging practices are available in the classroom, because of linguistic diversity, choosing one or two languages for teaching Maths and NSTech in linguistically diverse schools often overlooks learners’ linguistic diversity and further marginalises unpopular languages within the school setting. This led some educators to resort to monolingual practices, using English exclusively in their teaching. These practices indicate tensions between multilingual education policies that promote African languages and the practical pressures teachers experience in English-dominant educational environments. From a translanguaging perspective, such narratives reveal how monoglossic assumptions continue to influence educators’ understandings of language use and learner participation in township schools. This means that the inclusion criteria used for selecting languages for implementing MTbBE in multilingual township schools reveal what Makoe (2014) referred to as the socialisation of multilingual learners into homogeneous practices, as only one language is selected in each school, despite learners being linguistically diverse, under the assumption that a dominant language should carry out the implementation.

The results from this study also reveal that the inclusion criteria used in multilingual township schools often lead to the infringement of other learners’ rights to receive education in the language of their choice, as stated in the South African Bill of Rights (Republic of South Africa 1996), more especially in instances where learners do not share the identical linguistic repertoires as their educators. For instance, some educators reported that isiZulu was widely used in code-switching because it is most spoken in the community and at the school, suggesting that dominant languages often overpower non-dominant languages in multilingual classrooms, leading to linguistic diglossia. The problem with code-switching in this case is that code-switching practice excluded migrants and other learners whose first language was not isiZulu (Sibanda & Tshehla 2025) – giving the impression that other learners’ HLs did not have a place in the classroom, since they do not speak the languages used in the school for the implementation project. In contrast, the dominant languages of the society are used for the implementation project (Phajane 2022). Because of linguistic diversity in township schools and the inclusion criteria used to select languages for teaching Maths or NSTech, this study found that some learners are indeed learning in a language that is not their mother tongue, which defeats the purpose of MTbBE.

Although some educators mentioned translation and code-switching as strategies they use to teach complex concepts, other educators reported that, because of linguistic diversity in their classes, it did not accommodate all learners, which aligns with the literature, which states that oral translation in a multilingual classroom accommodates learners who share the same language as the teacher (Masunungure & Modiba 2023), meaning there is a great need for the DBE to provide translation services in schools and ongoing training for educators on how to embrace linguistic diversity in the classroom through multilingual pedagogies (Masunungure & Modiba 2023).

Educators in this study also reported that, prior to implementing MTbBE, there was no clear guidance on choosing the languages to teach Maths and NSTech in, even during the workshops they attended. Instead, they saw that the documents arriving in their schools had Maths and NSTech curricula written in isiZulu, meaning that indeed there is a lack of clear guidelines in implementing other policies (Mashige et al. 2019). This study revealed a need for educators to receive proper training to meet policy expectations and address issues related to inclusion, diversity and equity (Masunungure & Maguvhe 2024).

Data from document analysis revealed that although MTbBE is a bilingual approach, some documents, such as worksheets, were monolingual. Educators had to spend excessive time translating in class for learners who did not understand or speak the language in which the worksheets were presented. In other instances, they had to rectify language errors in those worksheets. This confirms the literature on language policy and linguistic landscape that although classes are highly multilingual, you hardly find multilingual documents that include three or more languages (Kretzer & Kaschula 2021). Similarly, Khanyile (2021) noted that implementing mother tongue instruction in SA is challenging because of limited resources and a lack of vocabulary in indigenous languages, as these languages lack registers for academic purposes across subjects such as maths, science, industry and commerce, which indicates the need to develop terminology and materials in indigenous languages (Sibanda & Tshehla 2025).

This study explored narratives on inclusion criteria within MTbBE implementation in selected township schools. Therefore, the findings should not be understood as neutral or objective representations of MTbBE implementation, but rather as contextually situated, shaped by multilingual classroom realities, institutional expectations and broader language ideologies surrounding the status of English and African languages. As participants were limited, they were sampled in a context-specific environment, which shaped their beliefs on MTbBE implementation. Nevertheless, this study revealed both the potential benefits and practical challenges associated with implementing MTbBE in linguistically diverse township classrooms. While the implementation of MTbBE is intended to promote epistemic access and equity, educators revealed that in other instances, like in multilingual township schools, it tends to marginalise unpopular languages, since they are hardly included in schools to be used to teach either Maths or Science; instead, school inclusion criteria tend to rely on dominant languages, which means that tension exists between policy ambitions and classroom realities in SA (Jaspers 2015; Makupe & Machaba 2025). Even though policies aim to promote multilingualism, educators reported that, in multilingual classrooms, they often draw on monolingual practices because of their language ideologies about the prestige status of other languages.

The inclusion criteria used to select languages in implementing MTbBE are suitable for those speaking languages shared by educators and bad for others, since the dominant ideology that was used as a basis of inclusion criteria for MTbBE in township schools reflects how homogeneity in heterogeneous environments reproduces differences in epistemic access and the education system, since it creates possibilities for some learners and constraints for others. Furthermore, it makes the job of mathematics teachers complicated and challenging, as they are not competent in the learners’ diverse, multilingual indigenous African languages (Makupe & Machaba 2024). The inclusion criteria for MTbBE in multilingual township schools give a false impression that learners in township schools are a homogeneous group. This is not the case, since schooling offers different experiences for individuals depending on their linguistic backgrounds (Krause & Prinsloo 2016). Given the exploratory and context-specific nature of the study, the findings cannot be generalised to all township schools or educational settings. Nevertheless, the study contributes to ongoing discussions on multilingual education and provides insight into how teachers implement MTbBE and navigate language practices and learner inclusion within township schools.

This study also contributes to the field of education by providing empirically grounded insights into how some educators interpret inclusion criteria within MTbBE implementation in township schools. While existing research has extensively examined language-in-education policies and translanguaging practices, less attention has been given to how educators operationalise language inclusion decisions in linguistically diverse classrooms. By focusing on educators’ situated perspectives, this study highlights the intersection of policy intentions and classroom realities, particularly in township contexts where English dominance continues to shape educational practice, despite measures to promote multilingualism.

This study has several limitations, one of which is that it involved a limited number of participants from selected township schools, and their self-reported experiences and perceptions may have been influenced by personal beliefs and professional experiences, limiting the generalisability of the findings to other educational contexts. The researcher also focused on specific township school settings where multilingual practices may differ from those in rural or urban schools.