In response to cyclones, flooding, and political insecurity following the 2024 Mozambican general election, tens of thousands of Mozambicans fled across the border into Malawi, particularly into the Nsanje and Dedza districts (Delegation of the European Union to the Republic of Malawi [EU] 2026). Emergency responses included humanitarian assistance supported by the EU. This situation illustrates the scale and urgency of the displacement experienced by local communities and schools. The linguistic impact of these movements is equally evident. Mozambique is itself a highly multilingual society with more than 40 languages, including Sena, which is spoken in the areas affected by extreme weather events and displacement, alongside Portuguese (Translators without Borders 2019). Moreover, cross-border linguistic continuity is sustained through Sena varieties spoken on both sides of the Malawi–Mozambique border. For example, a Sena-Portuguese-speaking immigrant entering Nsanje would immediately be understood by local speakers when asking to borrow a dish: palato in Sena and prato in Portuguese. Miscommunication would arise only if Chichewa, Malawi’s national language, were the sole language of interaction, as dish translates as mbale, a word that bears no phonetic resemblance to either Sena or Portuguese. Similarly, an immigrant would be understood when asking for batata (Portuguese), bambaya (Sena), or mbatata (Chichewa), but not potato (English). These examples illustrate the transnational nature of everyday linguistic repertoires.

Borderland scholars across Africa (e.g. Ndhlovu 2013a, 2013b; Ndhlovu & Masuku 2025) have documented precisely this kind of linguistic dynamism, in which cross-border languages and communicative practices do not conform to political boundaries. These scholars challenge the assumptions underpinning national language policies that view languages as fixed and bounded entities. This has direct implications for literacy, as literacy practices, including reading and writing, are shaped not by a single language code but by repertoires that continually flow and reconfigure through mobility, community ties, and crisis contexts (Abdi 2025). The educational value of such languages and practices has long been recognised at the continental level. Obanya (1999, cited in Alidou et al. 2006) identified 16 shared languages spanning 34 African countries and spoken by well over 140 million people, figures that have undoubtedly grown since then. The United Nations Educational, Scientific and Cultural Organisation (UNESCO) designates these as inter-African languages: languages that predate colonial borders and constitute important resources for education and regional communication, even though they are routinely excluded from what officially counts as languages of wider communication. Sena belongs to this category. Spoken on both sides of the Malawi–Mozambique border, it functions as the shared language of everyday borderland life in Nsanje yet holds no recognised place in the educational systems of either country. This exclusion mirrors, at a local level, the broader continental pattern documented by scholars.

Most language policies in multilingual African contexts remain heavily influenced by Western models, with policy approaches largely borrowed from the Global North (Aroge & Makalela 2024; Ndhlovu 2013b). This orientation overlooks local realities and contributes to tensions between policy and practice, thereby undermining effective implementation because of the disconnect between policy frameworks and the actual conditions within schools. Reilly, ResCue and Chavula (2022) argued that such policies reflect a monoglossic view of language because they view languages as distinct and separable instead of recognising the fluid multilingual repertoires that learners bring into classrooms. Importantly, Reilly et al. (2022:83) found that ‘in direct contrast to the monolingual English-only policy, the language practices which were observed in the primary schools in the Mangochi and Nsanje districts were highly multilingual’.

Malawi’s language-in-education policy has historically been unstable, reflecting both political reforms and contested debates over national identity and educational access. Mchazime (2001) and Kayambazinthu (2004) traced this history, showing that early post-independence policies recognised English and Chichewa as the official and national languages, respectively. This left limited functional space for other local languages in formal education, even though at least a dozen local languages were actively spoken (Chimpololo 2010). During the mid-1990s, with the advent of multiparty democracy, policy shifted to allow the use of local languages in education. The intention was to align classroom instruction with learners’ home languages. However, inadequate preparation, learning materials, and implementation support undermined this policy. As a result, the education system continued to oscillate between competing objectives: promoting national languages while responding to the realities of a multilingual population (Chimpololo 2010). In 2014, Lucius Kanyumba, the then Minister of Education, announced at a press briefing that Malawi’s language-in-education policy had changed and that English would become the language of instruction from Standard 1 (Kaphwiyo & De Souza 2025). For this reason, the current verbal policy is mockingly known as ‘Kanyumba English-Only Policy’.

There was little evidence of systematic language planning. Language planning is essential for fostering social cohesion, inclusivity, and educational equity. It encompasses deliberate efforts by governments to influence the status, structure, and acquisition of languages in response to broader social needs (Cooper 1989). Even when multiple indigenous languages were promoted alongside Chichewa and English, policy intentions did not translate into effective classroom practice, largely because of resource constraints, limited teacher preparation and inadequate curricular support for multilingual instruction (Chiuye & Moyo 2008). Instead, Chichewa continued to serve as the primary medium of instruction in the early years of primary education, with English assuming dominance from Standard 5 onwards. A similar disjuncture has been documented south of the border. In their evaluation of literacy teaching in primary schools in Limpopo Province, South Africa, Reeves et al. (2008) found that although classroom discourse was de facto multilingual, with teachers routinely code-switching, learners were required to read, write, and be assessed monolingually in English from Grade 4. This gap between multilingual practice and monolingual literacy demands closely parallels the Malawian situation described in this article.

Studies have shown that learners, particularly in multilingual contexts, are seldom linguistically prepared for the abrupt shift from Chichewa to English as the language of instruction. Moreover, promoting a single national language within policy is problematic because it marginalises other languages. This concern was anticipated in the continental policy literature. The Association for the Development of Education in Africa (ADEA) and the UNESCO Institute for Education advocacy brief, which synthesised evidence from a 25-country review, recommended that African education systems promote mother-tongue-based multilingualism while cautioning that ‘colonial monolingualism is not replaced with African monolingualism’ (Ouane & Glanz 2010:6). Malawi’s elevation of Chichewa illustrates precisely this concern: a policy presented as a decolonial affirmation of an African language reproduces, for Sena-speaking borderland learners, the same exclusionary logic it claims to replace, namely instruction through a language that is not their own, rendering their actual linguistic repertoires invisible. Ouane and Glanz (2010) therefore argued for a gradual and all-inclusive approach to language choice in education instead of the either-or logic that positions one language against another. Earlier, Mchazime (2001) similarly argued that this mismatch limits meaningful participation, depresses achievement in Social Studies and other content areas, and encourages teacher-centred instruction while constraining learners’ active engagement.

Learners must attain certain proficiency thresholds in both their primary language and the language of instruction for bilingualism to be effective, particularly in terms of cognitive development. Heugh’s (2006) typology of language education models in Africa makes the structural problem clear. Malawi’s longstanding arrangement, in which Chichewa served as the medium of instruction until Standard 4 before English assumed that role, represents a transitional model that withdraws learners’ familiar language after only 4 years. This falls well short of the minimum of 6 years of mother-tongue instruction that continental evidence indicates is necessary, and even further short of the 6 years to 8 years generally required for learners to develop sufficient proficiency in a second language to learn successfully through it. According to the continental stock-taking study, such weal bilingual models leave learners with inadequate proficiency in the international language by the end of schooling, resulting in poor performance across the curriculum. The 2014 directive did not address this weakness; instead, it intensified it by shifting Malawi from a transitional model to a subtractive, straight-for-English model, the configuration associated with the poorest documented outcomes.

Developing cognitive skills and proficiency in one language before progressing to another enables learners to cope more effectively with academic demands, and this proficiency positively influences academic achievement (Chaudhary 2018). This approach aligns with multilingual education research, which identifies two intertwined challenges in contexts such as Malawi: (1) policy frameworks that fail to harness learners’ full linguistic repertoires for meaning-making; and (2) classroom pedagogies that view languages as discrete codes rather than as interconnected resources through which learners make sense of texts and contexts (Reilly et al. 2022). These challenges are not merely technical; they are also deeply socio-political because they marginalise linguistically minoritised learners whose everyday communicative practices are erased or devalued in schooling (Manthalu & Ngwira 2020).

This monoglossic orientation is neither new nor undocumented. More than two decades ago, a 25-country stock-taking study commissioned by UNESCO–ADEA (Alidou et al. 2006) concluded that education delivered through ex-colonial languages unfamiliar to learners results in systemic failure across the curriculum. The study did not stop at diagnosis; it advanced substantive recommendations for mother-tongue-based bilingual and multilingual models. The fact that the situation described in Nsanje persists more than two decades after this continental diagnosis confirms that the linguistic exclusion of borderland learners is not an unforeseen consequence of recent displacement, but rather the predictable outcome of a long-recognised and persistently neglected policy failure.

Conventional approaches to literacy education are deeply shaped by what has been described as a monoglossic ideology of language, an epistemological stance that views languages as bounded, stable, and nameable entities that can be cleanly separated, standardised, and sequentially acquired (Reilly et al. 2022). Within this worldview, reading and writing are conceptualised as technical skills that develop through mastery of a single authorised language of schooling, while other languages are positioned as obstacles, transitional scaffolds or, at best, extracurricular resources. Such assumptions continue to underpin language-in-education policies across much of Africa, despite decades of sociolinguistic evidence demonstrating that multilingualism is the norm rather than the exception (eds. Erling et al. 2022; eds. Kretzer & Kaschula 2022; Maalim 2014; Schroeder, Mercado & Trudell 2022).

Philosophically, this fixation on linguistic stability reflects a deeper modernist impulse. It reveals a desire to govern complexity through order, classification, and containment. Languages are imagined as objects that can be mapped onto countries, curricula, and assessment models, while literacy is framed as a stand-alone technology detached from social life. However, in contexts shaped by displacement, precarity, and mobility, such as Nsanje, this epistemology becomes untenable. Ehret and Rowsell (2021) argued that literacy is not acquired under conditions of stability but forged under conditions of movement, uncertainty, and contact. Reading and writing are therefore not merely school practices but also survival practices in border crossings, humanitarian encounters, and negotiations of belonging.

Ndhlovu (2013a) argued that languages derive their value from the social capital generated through social networks and their practical functions within communities. In the context of this study, languages have become essential for communication within communities and across schools. This has important implications for educational practice. If these languages function as communicative resources beyond the classroom, then excluding them from pedagogy represents both a missed educational opportunity and an epistemological contradiction. When learners’ linguistic repertoires are subordinated to colonial languages, their languages and their capacity to draw on lived realities in learning and meaning-making are devalued. These repertoires embody community knowledge and distinctive networks of meaning-making that should likewise be recognised as resources for literacy development.

From this perspective, the question is not whether immigrant learners from Mozambique possess the ‘right language’ for literacy learning, but how schooling can recognise, embrace and pedagogically mobilise the repertoires they already use to make meaning across borders. To theorise literacy under conditions of borderland mobility, we introduce the metaphor of the braided river. We do not claim to have invented this metaphor, as it has been used widely in other contexts (e.g. Curtis et al. 2013; Keyes, Ross & Chan 2012; Prentice 2025). Rather, we contextualise it within multilingual literacies for immigrant learners. Unlike a single-channel river that flows predictably within fixed banks, a braided river consists of multiple interweaving channels that diverge, converge, and shift over time in response to environmental pressures such as flooding, sedimentation, and changing terrain. Stability in such rivers is achieved not through rigidity but through adaptive movement. Applied to language and literacy, the braided river metaphor offers several important insights.

First, it captures the plurality of linguistic currents that characterise Nsanje’s classrooms: Sena varieties shared across the Malawi–Mozambique border, Portuguese-influenced registers carried by Mozambican learners, Chichewa as a national language of schooling, and English as a gatekeeping language for academic progression. These are not separate streams flowing in isolation but interdependent channels that continually shape one another through constant contact. Second, the metaphor highlights the role of disruptive events – floods, cyclones, and election-related violence – as forces that continually reconfigure linguistic practices. Just as flooding reshapes a river’s channels, displacement reshapes learners’ literacy trajectories. In turn, these generate new communicative needs and hybrid practices that schools are seldom prepared to accommodate. Third, the braided river challenges policy imaginaries that seek to ‘canalise’ language use in classrooms. Language-in-education policies that insist on a single medium of instruction attempt to force a braided river into a narrow channel (cf. Curtis et al. 2013). The result is not clarity but erosion, as learners’ meaning-making resources are stripped away, participation narrows, and literacy becomes an exclusionary rather than an inclusive practice. The braided river metaphor therefore serves as a conceptual bridge between philosophical critique and pedagogical design. It invites educators and policymakers alike to ask not how linguistic multiplicity can be eliminated, but how teaching can occur within it.

Building on the braided river metaphor, we propose the BMF for understanding and designing literacy education in contexts of cross-border mobility and displacement. This framework builds on Heugh’s (2021) notion of transknowledging. Because the use of multiple languages engages both linguistic and epistemic dimensions, and because knowledge developed in one language may not have direct equivalents in another, multilingual pedagogies enable the exchange of diverse knowledge systems, a process that Heugh (2021:43) identified as transknowledging: ‘the process of knowledge exchange and production’. The systematic use of multiple languages in classroom contexts therefore promotes inclusivity and social cohesion among learners while enhancing their capacity for meaning-making. This begins by normalising the use of learners’ primary languages and broader linguistic repertoires. Second, it requires the strategic deployment of multilingual practices, involving deliberate decisions about when and why different languages should be drawn upon. Recognising multiple languages also requires teachers to possess the knowledge and skills needed to use learners’ repertoires effectively as pedagogical resources. This further entails recognising that each language embodies distinctive knowledge systems. These knowledge systems can then complement one another in constructing new understandings. In this way, transknowledging maximises knowledge exchange while facilitating literacy development and recognising the significance of language, knowledge, and inclusive meaning-making.

The BMF argues that place, mobility, and crisis are not peripheral factors but constitutive conditions of meaning-making. Borderlands are not merely multilingual spaces. As Ndhlovu (2013a) argued, they are politically charged zones in which languages, identities, and educational futures are continually renegotiated. From this perspective, the BMF is underpinned by five propositions:

Collectively, the braided river metaphor and the BMF provide a conceptual framework for multilingual literacy education, as illustrated in Figure 2.

The braided river metaphor captures how linguistic currents (Sena varieties, Portuguese-inflected registers, Chichewa, and English) diverge and recombine across social sites and over time. It also demonstrates how disruptive events, such as floods and elections, reconfigure channels, deposit new linguistic practices and require adaptive literacies. The BMF builds on multilingualism and borderland discourse to position mobility, crisis responsiveness, and repertoire entanglement as central analytical perspectives for literacy research in cross-border contexts. The philosophical and theoretical reframing offered by the braided river metaphor and the BMF lays the foundation for a pedagogical shift from monolingual transmission models of literacy towards multilingual, repertoire-based approaches to reading and writing. Instead of asking how immigrant learners can be assimilated into existing language regimes, the more pertinent question becomes how classrooms can be reimagined as linguistically porous spaces that reflect the realities of borderland life.

Under monolingual policy regimes, literacy instruction typically aims to develop proficiency in a single authorised language and views multilingual resources as noise within the system (Eisenchlas, Schalley & Guillemin 2015). In contrast, multilingualism-informed pedagogy reframes literacy goals towards meaning-making and epistemic access through and across multiple knowledge and language systems. This shift aligns with sociocultural conceptions of literacy as a practice rather than merely a skill. Research on multilingual education demonstrates that when learners’ full linguistic repertoires are recognised as resources for learning, participation, comprehension, and identity affirmation, all these aspects are enhanced (Carrim & Nkomo 2023). For example, in South Africa, multilingual strategies such as terminology development (Nkomo & Madiba 2011), translation (Mawonga, Maseko & Nkomo 2014), the use of learners’ linguistic repertoires, and translanguaging (Lepheana 2021) have been shown to enhance multilingual learners’ literacy development by enabling home language use, translation, and flexible language practices in support of deeper meaning-making. In Nsanje, such a reframing would mean that reading lessons do not merely focus on decoding in Chichewa or English. Instead, reading becomes a process of meaning negotiation in which learners draw on their existing Sena repertoires and Portuguese-influenced linguistic patterns to make sense of texts presented in any language of instruction.

Multilingual pedagogy encourages the use of text sets that draw on multiple languages. In the context of this study, these could include oral stories narrated in Sena, short written texts in Portuguese, picture books with bilingual captions, and multilingual glossaries. This, in turn, requires teachers to understand how to respond to and implement multilingual pedagogies effectively. Research shows that such multimodal and multilingual learning environments promote learner engagement and a sense of ownership of literacy tasks (Tai 2022). Teachers can structure reading tasks that invite learners to compare, contrast, and synthesise meanings across languages. Such tasks both acknowledge learners’ linguistic repertoires and strengthen literacy across languages. For example, at an appropriate school level (e.g. Standard 5), learners could read a simple informational text on flood safety in Chichewa, discuss the same topic orally in Sena, and then write reflections in English informed by both discussions. Similarly, at a more advanced primary school level (e.g. Standard 8), learners could develop multilingual reading dictionaries that link key vocabulary across Sena, Portuguese, Chichewa, and English. Such activities can promote metalinguistic awareness and strategic language use.

Traditional writing instruction views writing as the production of standardised text in a single target language (Kubota 2022). By contrast, Kaufhold (2018:1) found that ‘students’ linguistic ideologies and their experiences can enable or restrict their capacity to draw on their varied repertoires’. Multilingual approaches therefore conceptualise writing as repertoire assembly, in which learners draw on whatever linguistic and semiotic resources they possess to construct meaning. In practice, this may involve bilingual journals in which learners initially draft entries in Sena or Portuguese before co-constructing versions in the language of assessment (e.g. English) with teacher support. It may also involve multilingual, multimodal compositions that combine written text, drawings, and oral recordings. This approach is supported by multiliteracy research, which emphasises the value of diverse semiotic resources in literacy learning (see Pennycook 2017). Such activities validate learners’ linguistic identities while supporting the development of critical literacy that attends to how and why meaning is constructed across linguistic repertoires.

Multilingual pedagogy should not be left to spontaneity. Rather, it should be intentionally structured through pedagogical design (Faltis 2020). Depending on the context, this may involve a combination of multilingual strategies. Such an approach can be organised around three interconnected principles: recognising multilingualism as the norm in schooling contexts by creating classroom environments in which learners feel safe and empowered to use their full linguistic repertoires for learning; strategically drawing on learners’ languages to enhance literacy through translanguaging, code-switching, translation, and other pedagogical strategies that facilitate epistemic exchange; and promoting dialogue among multiple knowledge systems, thereby enabling transknowledging. This requires teacher modelling, explicit negotiation of which languages are available for use, and the establishment of classroom norms that position all languages as accepted resources for learning (Tai 2022). The Languaging for Learning (L4L) approach supports teachers in designing learning sequences in which translanguaging promotes simultaneous access to content, literacy skills, engagement, and comprehension (Hendricks & Xeketwana 2024). Such an approach can also be used to build teachers’ capacity to work effectively in multilingual environments.

Of course, when discussing spontaneous and pedagogical translanguaging, it is important to acknowledge the intellectual lineage of these concepts. While we draw on Faltis’s (2020) important distinction between spontaneous multilingual language practices and intentionally designed pedagogical translanguaging, similar ideas were articulated earlier by Cenoz and Gorter (2017), who drew on Bernstein’s (1999) distinction between horizontal and vertical discourse. As they (Cenoz & Gorter 2017) explain:

This concern is echoed by Jaspers (2018), who argues that powerful languages tend to remain powerful while minority languages remain marginalised, even when translanguaging occurs. This suggests that spontaneous multilingual practices alone are insufficient; they must be accompanied by deliberate policy, institutional structures, and cultural affirmation that actively support minority and cross-border languages.

This perspective is particularly relevant to our argument because it cautions against assuming that the spontaneous or horizontal use of multilingual repertoires in everyday communication is, on its own, sufficient to support curriculum learning or equitable educational outcomes. As Faltis (2020) and Heugh (2021) both emphasised, multilingual practices need to be systematically incorporated into pedagogical design so that learners can access disciplinary knowledge, develop academic literacies, and succeed in high-stakes assessments. In the context of borderland multilingualism, this means that recognising learners’ fluid linguistic repertoires is a necessary starting point, but educational inclusion ultimately requires deliberate pedagogical structures that transform these repertoires into resources for sustained learning, transknowledging and academic achievement.

Assessment in monolingual regimes usually penalises multilingual learners whose written or oral responses do not conform to strict language norms (Gorter & Cenoz 2017). A multilingualism-informed approach to assessment reframes assessment tasks as meaningful communicative performances that may draw on multiple languages. In the context of this study, possible approaches could include portfolio assessment, in which learners compile multilingual texts (such as diaries, translations, and collaborative compositions) over time to demonstrate literacy development across modes and languages. It could also include performance-based assessment, whereby learners present on a topic (e.g. community life in Nsanje and neighbouring Mozambican villages) using a combination of languages, visuals, and oral explanations, with assessment focusing on comprehension and communicative effectiveness rather than strict linguistic conformity. Bailey et al. (2025) observed that such equity-oriented practices can help reveal linguistic competencies that monolingual assessments often obscure.

For multilingual pedagogy to succeed, teachers’ beliefs and professional preparation must also evolve. Research indicates that long-standing monolingual ideologies, particularly the belief that standardised languages are inherently superior for learning, present significant barriers to implementation (see Chang-Bacon 2021). Teacher development should therefore include critical reflection on language ideologies to support teachers in examining assumptions about language hierarchies and recognising the value of multilingual repertoires (McKinney & Guzula 2024). Capacity building should also include teacher education in bilingualism, multilingualism, and biliteracy. Within an inclusive education context, partnerships with families and communities are equally important, recognising that immigrant households possess valuable linguistic and cultural knowledge that can enrich classroom literacy practices. Such professional development and community engagement are essential components of sustainable pedagogical transformation.

The pedagogical implications of multilingualism for reading and writing in borderland classrooms highlight tensions between what teachers can do in their classrooms and what existing policy requires them to do. In contexts such as Nsanje, where immigrant learners and local communities share multilingual repertoires that defy neat linguistic boundaries, the insistence on a monolingual policy framework represents a structural barrier to inclusive literacy education. Transforming this landscape requires not only classroom-level change but also language policy reform that creates space for multilingualism as a recognised educational goal and practice.

Malawi’s current language-in-education policy reflects a monoglossic conceptualisation of language, in which a single medium of instruction is designated for formal schooling (Kaphwiyo & De Souza 2025). Following the Kanyumba English-Only Policy in 2014 (which was announced orally and has never been formalised through regulations or review mechanisms), English has been designated as the language of instruction from the early years of primary schooling, with Chichewa playing a reduced role, resulting in a subtractive bilingual model. Although this policy change is justified in terms of perceived global relevance and language proficiency, criticism of it is well documented. Trudell’s (2021) critique of globalised curricula is particularly relevant here, highlighting how local knowledge and languages become undervalued and marginalised. Scholars argue that the policy disadvantages learners from rural and peri-urban communities where English is not widely used as a second or additional language. As Manthalu and Ngwira (2020) observe:

Thus, elevating English as the exclusive language of instruction excludes local epistemologies and cultural expressions. In turn, this constrains inclusive education and marginalises local identities in favour of colonial linguistic norms.

Despite the official monolingual policy, classroom practices are inherently multilingual, with teachers and learners drawing on diverse linguistic repertoires to make meaning, albeit largely informally because policy does not permit such practices. Indeed, Kaphwiyo and De Souza (2025) observed that teachers tend to adhere strictly to the Kanyumba Policy primarily during official inspections because they themselves, as members of the local linguistic ecology, often struggle to teach exclusively through English. This illustrates a policy environment that fails to recognise multilingual repertoires as pedagogical resources rather than obstacles. Given this disparity between policy and classroom realities, several high-leverage policy directions could inform a reformed language-in-education framework for Malawi, particularly in borderland districts such as Nsanje.