Introduction

Approaches to language learning in postcolonial Africa have been based largely on European and North American pedagogical theories inherited from the colonial period and have focused on the implementation of monolingual instruction in English, French, and Portuguese (Wolff, 2017; Heugh, 2021). In recent decades, however, there has been a shift towards developing transformative forms of education that reflect African cultures, values, and languages and that take into account sociocultural considerations as well as educational needs (Alidou et al., 2006), with scholars arguing for the value of multilingualism and multilingual education based on the L1 (see e.g. Bamgbose, 2000; Chimbutane, 2011, 2018; Cummins, 2005; Djité, 2008; García & Kleifgen, 2010; Ouane & Glanz, 2010; World Bank, 2021). As such, in African countries there has been an increase in educational initiatives that embed local languages and cultures into the curriculum and many countries have developed bilingual1 educational models in order to meet local learning needs while maintaining the use of the dominant language as the language of global communication (Diallo, 2011). Nonetheless, there are still challenges around the implementation and adoption of bilingual education, as this article will discuss.

This article draws on the case of Cabo Verde, where Portuguese is the only official language of schooling, despite the Cabo Verdean language (CVL)2 being the first language of virtually the entire population. This mismatch between the language of instruction and the language of students presents a challenge for the education system (Batalha & Carling, 2008). Such linguistic challenges are heightened for children from the most rural and disadvantaged backgrounds who, unlike those from urban middle class and elite Cabo Verdean society, have limited exposure to Portuguese. As a study that forms part of a broader research project, this research examines a grassroots bilingual initiative to introduce CVL alongside Portuguese as a medium of instruction in two primary schools. It looks specifically at the language ideologies that emerged in interviews with selected teachers, activists, and politicians regarding the introduction of bilingual education and questions why, despite the recognised evidence of the benefits of bilingualism, there is still resistance to implementing bilingual education programmes in Cabo Verde.

In order to analyse the tensions surrounding the medium of instruction in a postcolonial setting such as Cabo Verde, we adopt and expand on the concept of the “access paradox”, which has been defined by Janks as follows:

If you provide more people with access to the dominant variety of the dominant language, you contribute to perpetuating and increasing its dominance. If, on the other hand, you deny students access, you perpetuate their marginalisation in a society that continues to recognise this language as a mark of distinction. You also deny them access to the extensive resources available in that language; resources which have developed as a consequence of the language’s dominance. (2004, p. 1)

Applying the access paradox to the Cabo Verdean context, the monolingual Portuguese curriculum perpetuates the dominance of Portuguese and therefore naturalises its power and devalues CVL. In contrast, excluding Portuguese from the curriculum excludes students from a variety that would afford them linguistic capital, and restricts them to the communities and linguistic markets where CVL is spoken, arguably perpetuating marginalisation. We argue, however, that this introductory explanation is somewhat reductionist, and that by bringing insights provided by bilingual education debates, we can expand on the concept of the access paradox, drawing on the case of Cabo Verde specifically to illustrate the complexities of bilingual education in postcolonial contexts more broadly. Perhaps most significantly, in this article we identify how the access paradox is firmly rooted in language ideologies. In order to understand how the access paradox operates in a postcolonial context, it is essential to understand the complexities of the language ideologies that sustain resistance to bilingual education.

Language Ideologies

This study draws on language ideologies as an analytical framework to unpack the ways in which speakers view and understand language (Cavanaugh, 2011, p. 46). In one of the earliest explorations of the concept, Silverstein (1979) defines language ideologies as “sets of beliefs about language articulated by users as a rationalisation or justification of perceived structure and use” (p. 193). Here, however, we draw on Kroskrity (2010, p. 192) and define language ideologies as “beliefs, feelings and conceptions about language structure and use which often index the political economic interests of individual speakers, ethnic and other interest groups, and nation states”, thus highlighting the sociocultural and sociopolitical nature of language ideologies. Moreover, we wish to highlight that language ideologies are about more than just language, in that “they envision and enact ties of language to identity, to aesthetics, to morality and to epistemology” (Schieffelin, Woolard, & Kroskrity, 1998, p. 3). As Horner and Weber (2017) note, language ideologies are usually “imbued with vested interests and can play a role in group membership and boundary negotiation, as well as social inclusion and exclusion” and can therefore be considered as “the cultural systems of ideas and feelings, norms and values, which inform the way people think about languages” (p. 20). This is particularly relevant for the present study, as it aims to understand how bilingual education is related to the values, both emotional and instrumental, associated with the languages in question.

Dominant ideologies, as “hegemonic beliefs and feelings about language that both reflect and serve the interests of groups with social, economic and/or political power” (Martínez, 2013, p. 278), are integral to the reproduction of social and power structures (Jaffe, 2008). In this article, we identify three interrelated language ideologies that form part of a broader dominant ideological framework: standard language ideology, ideologies about language hierarchies, and the ideology of one nation-one language. While for the purposes of this discussion, these are presented as three separate ideologies, their interconnected nature must be taken into account, and this is indeed reflected in how they emerge in this study. First, standard language ideology can be understood as “a bias toward an abstract, idealised homogenous language which is imposed and maintained by dominant institutions” (Lippi-Green, 1997, p. 64). Notably, it is sociopolitical factors that lead to a particular variety becoming the standard, rather than any intrinsic superiority of the variety (Horner & Weber, 2017). Within this framework, language is essentialised and positioned as an unchanging system (Boudreau & Dubois, 2007; Jaffe, 2008). Second, ideologies about language hierarchies centre on a belief that linguistic varieties can be classified in a hierarchical structure, as either languages or dialects, for example. Within this ideological framework, languages are usually “looked upon as superior to ‘dialects’ and, additionally, certain languages [are] given a higher status as the ‘national’ or ‘official’ language of the state or community” (Horner & Weber, 2017, p. 11). Third, the one nation-one language ideology rejects linguistic hybridity and plurality in favour of a model which equates language with territory and with national identity in a Herderian territory-culture-language triad (Blommaert & Verschueren, 1998). Thus, in this study, we examine how this interconnected, dominant ideological framework emerges in the discourses of participants whereby (standard) languages are viewed as clearly delineated separate entities that are maintained in a hierarchical relationship across all domains. Furthermore, we see that it is the association with the nation or the former colony that provides each language its particular meaning and perceived value.

Bilingual Education

Africa

The United Nations Sustainable Development Goal number 4 aims to “ensure inclusive and equitable quality education and promote lifelong learning opportunities for all” by 2030. Part of ensuring inclusive and quality education includes addressing the question of medium of instruction (UNESCO, 2016). These questions are not new, and date back to the UN’s landmark publication in 1953, which emphasised the importance of L1-based education for effective learning. In this article, we follow Chimbutane (2011), Cummins (2005), Heugh, Siegrühn, and Plüddemann (1995), and a large body of literature that contends that languages in the classroom can “nurture” each other. It has been shown that policies that foster multilingualism can enable people to contribute in more creative ways to the economy (Djité, 2008) and L1-based instruction has been shown to promote active learning and meaningful participation in the classroom by raising students’ confidence and self-esteem (Lawrence Gordon & Harvey, 2018; Ouane & Glanz, 2010; UNESCO, 2016), with research pointing to how, when hegemonic regimes are lifted, students can change some of the established language hierarchies in the classroom and pave the way for innovation in plurilingual pedagogies (Kerfoot & Bello-Nonjengele, 2016). It is important to note that the benefits of including L1-based instruction in the education system are not limited to language proficiency: results from pilot studies in schools across Africa show that L1-based instruction leads to better academic performance across subject areas (UNESCO, 2016) and language has been found to be a key factor in preventing educational repetition, failure, and drop-out rates (Benson, 2014, 2021).

In the present-day knowledge economy, it is high levels of literacy, “regardless of the number of languages involved, that characterise the most successful polities” (Alexander, 2007, p. 15). In other words, developing key literacy skills in early years (regardless of the language involved) is most important for effective learning. Moreover, teaching and examination in languages that children do not speak at home has been shown to hinder early acquisition of crucial literacy skills (UNESCO, 2016). Thus, bilingual initiatives aim to overcome the access paradox by ensuring the high levels of literacy (in two or more languages) needed for participation in the knowledge economy while allowing students to access the economic benefits provided by the dominant language (Janks, 2004). Crucially, the educational strategies that advocate for the inclusion of local languages and teaching through L1 do not propose rejection of the dominant language/language of the former colonial power. Rather, L1-based multilingual education improves learning across subject areas (Alidou et al., 2006).

Despite the documented benefits of bilingual education, the “language factor” (see Wolff, 2006) and the importance of including African languages across all areas of development—especially in education—have not received sufficient attention (Chimbutane, 2017). In most African countries, the language of instruction in school continues to be that of the former colonial power. Of the 17 countries in West Africa, all but one have maintained the former colonial language as their official language (the exception being Mauritania, which has Arabic as its official language) (Bamgbose, 2000). Furthermore, Africa is the only continent where the majority of children start school in a language other than their home language (Ouane & Glanz, 2010). Many of the recently established education models in Africa continue to be subtractive, with proficiency in L2 (usually the former colonial language) as the predominant or exclusive end goal (Heugh, 2011b). While anglophone African countries have tended to focus on transition or “early exit” education models which include some instruction in L1, lusophone and francophone African countries are usually characterised by monolingual submersion models inherited from the colonial period (Wolff, 2006), and this is indeed what we see in Cabo Verde. Heugh (2011b) notes that Mozambique is moving towards transition models (i.e. three years of education in L1 followed by Portuguese), but the model is still subtractive, with monolingual Portuguese instruction as the ultimate goal. The intersection between Africa’s colonial past and the challenges of modern-day globalisation are key factors in the resistance from policymakers and stakeholders (such as teachers, parents, and students) to adopting additive bilingual education policies in low- and middle-income contexts (Kananu Kiramba, 2018; Ouane & Glanz, 2010). These challenges have been explored extensively by scholars (see e.g. Alidou et al., 2006; Antia, 2021; Banda, 2000; Benson, 2021; Diallo, 2011; Probyn, 2009), and include issues relating to insufficient resources allocated to teacher training and the development of relevant materials, scarcity of academic textbooks in African languages, and broader concerns about cost implications of implementing bilingual education (for a thoughtful and nuanced analysis of the costs associated with implementing bilingual education in low- and middle-income contexts, see Heugh, 2011a).

Cabo Verde

In 1975, while the Partido Africano para a Independência de Guiné e Cabo Verde [African Party for the Independence of Guinea and Cabo Verde or PAIGC]3 was in power, the first post-independence education system for Cabo Verde (and Guinea-Bissau) was created. The renowned Brazilian pedagogue, Paulo Freire, played a key role in this process, and drawing on his theories of liberating pedagogy, he contributed to the development of literacy campaigns. Nonetheless, as part of these campaigns, Freire had to renounce the use of CVL, as policymakers argued that Portuguese had greater value internationally, and they considered the prioritisation of Portuguese to be politically beneficial (Rodríguez, 2017). Freire went on to regret this choice but felt at the time that he should not interfere in such strategic policy decisions (Freire & Macedo, 1989). Thus, while the end of the colonial regime in 1975 brought about educational reform and the creation of universal public education (which since 2006 extends to universal tertiary education), this has never included the use of CVL.

As in other minority language contexts, the standardisation and officialisation of CVL has been a recurrent theme in social and political debates since independence. Notably, as far back as 1979, the Fórum Internacional de Valorização do Crioulo [International Forum for the Valorisation of Crioulo], held in Mindelo, recommended that greater value be placed on CVL and made recommendations for the inclusion of the language in the school system, especially in its written form. Alphabets and grammars have been developed for CVL (e.g. Veiga, 1995); however, while the officialisation of the now widely used ALUPEC4 alphabet was trialled in 1998, it was not legally approved until 2009. Furthermore, many recommendations have been made for the inclusion of CVL in the education system. For example, as recently as April 2021, a group of more than 200 people linked to education and research submitted to the President of the Republic a petition for a change in language policy in the country, defending, among other measures, the implementation of bilingual education. Despite these developments, the education system remains monolingual in Portuguese. In short, arguably due to lack of explicit political support, we now see a situation where CVL holds a subordinate position to Portuguese in the language hierarchy even though it has never explicitly been denied equal footing with Portuguese.

The first bilingual education programmes to introduce CVL in schools did not emerge on the Cabo Verdean archipelago, but rather on the east coast of the United States, home of the largest diaspora community since 1968 (particularly in New England, where much of the Cabo Verdean diaspora is concentrated) (Tavares, 2020). In Portugal, a bilingual programme was offered for children at the Vale da Amoreira (Setúbal) school from the 2008/9 to the 2011/12 school year, with funding from the Calouste Gulbenkian Foundation. The “one year group-two languages” model was promoted as part of a study conducted by the Theoretical and Computational Linguistics Institute (ILTEC), which looked at linguistic diversity in Portuguese schools and situated CVL as the second most widely used language in the Greater Lisbon area (second only to Portuguese). This programme introduced a daily teaching hour in CVL to one primary year group with twenty-two children, nine of whom had previous contact exclusively with Portuguese. The remaining children were of Cabo Verdean (11), Guinean (1), and Angolan (1) descent. The participating year group demonstrated academic results superior to that of the control group, as well as less prejudice and more positive attitudes towards linguistic diversity. The children of Cabo Verdean origin also reactivated or broadened their knowledge and practice of their L1 (Mateus, 2011).

As for Cabo Verde itself, the first bilingual programme was introduced in the 2013/14 school year as part of a doctoral research project led by Ana Josefa Cardoso. The bilingual education initiative, the only one of its kind in Cabo Verde to date, had the authorisation and support of the government, although it was not part of formal educational policy. The initiative was implemented in two class groups for a period of six years and included provision for the training of teachers. In February 2015, a workshop in Praia—organised by the Ministry of Education and inaugurated by the then prime minister and minister for education—was held to discuss the initial successes of the programme. Two months later, the government approved a resolution (32/2015) which advocated for the promotion of CVL and highlighted that bilingual education was more suited to the context of the country. The Partido Africano da Independência de Cabo Verde [African Party for the Independence of Cabo Verde, PAICV] went on to lose the 2016 elections, and the recommendations from resolution 32/2015 never materialised; while the results of the bilingual pilot project pointed to benefits in terms of overall school performance and in language skills specifically, the project was not pursued past the pilot stage.

Methodology

Data were collected in 2018 through interviews with teachers and other agents (politicians, educational administrators, and activists) who were involved, directly or indirectly, in Cabo Verde’s 2013–2014 bilingual pilot initiative. This programme was supposed to last from the first to sixth year of primary, completing the full educational cycle. However, at the time of our data collection, the completion of the programme was uncertain, as the government which had approved and supported it (the PAICV) was replaced by the Movimento para Democracia [Movement for Democracy, MpD] in 2016. Government officials and teachers spoke of either a permanent or temporary suspension of formal authorisation.5 This context is important for three reasons: (1) the latter administration ceased to fully fund and authorise the bilingual project that was implemented by their predecessors; (2) the platform of the MpD does not, explicitly, include any anti-bilingual or anti-CVL rhetoric; and (3) the reasons for the decision were not publicly stated, giving rise to a multiplicity of interpretations.

Both participating schools were located on the Island of Santiago: one in a mountainous rural region of the interior (Flamengos, in the municipality of San Miguel), and the other in a large urban setting—the capital city of Praia. In both cases, the schools served a low-income population. We conducted both semi-structured individual and discussion group interviews in both locations. All participants were fluent in Portuguese and were asked to speak in Portuguese to facilitate the understanding of the two researchers who did not understand CVL. The presence of a third researcher who does understand CVL meant that participants were able to draw upon this linguistic resource occasionally. We provide here a general overview of participant profiles rather than more extensive individualised profiles, in order to preserve the anonymity of actors who, given the relatively small scope of the pilot project, could be easily identified.

The individual interviews were conducted with two male teachers who taught on the bilingual project; a female former education official who authorised, supported, and supervised the pilot project in one of the schools during the mandate of the PAICV; a male current education official working under the current (MpD) mandate; a CVL language activist who is well-known for his long history of advocating for CVL language rights in general, and defence of bilingual education in particular; and Ana Josefa Cardoso, on whose doctoral research the bilingual project was based. A further sixteen people participated in the two discussion group interviews, ten in the urban location and six in the rural: these included eleven female and three male teachers at both primary and secondary level who knew of but had not directly participated in the bilingual project. Two women who worked in social and administrative services associated with the schools also participated, although they did not speak during the interviews. During the data generation phase, we recorded field notes of the key trends as they emerged, as well as issues that we wished to further clarify or contrast with different people’s perspectives. These notes comprised a preliminary analysis that led to the creation of initial coding categories, so that once the interviews were transcribed we used these categories to code data using Atlas-ti software. Interviews were first analysed individually (vertically) and then as a group (horizontally). In the final analysis, we explored the ways in which participants drew upon broader discourses to provide their interpretations and support their own arguments about the value and viability of the bilingual pilot project. We used critical discourse analysis (CDA), which is “fundamentally concerned with analysing opaque as well as transparent structural relationships of dominance, discrimination, power and control as manifested in language” (Wodak, 2001, p. 3), as a tool to analyse how the arguments constructed by the participants interacted with broader language ideologies that serve to maintain the monolingual status quo of schooling, especially in postcolonial contexts where the colonial language maintains its higher symbolic value long after political independence has been secured. The data set in this study is understood as collaborative, co-produced, and situational (Talmy, 2010). Our aim is not to uncover what people really think; instead, we consider interviews as “situated performances” (Heller, 2008, p. 256) and see language as a form of social action which can provide insights into people’s ideological positionings (Moyer, 2008).

Data Analysis

Living in CVL, Schooling in Portuguese

There was an especially strong tendency for participants to highlight the divide between school and life outside the classroom, not surprising in a diglossic society like Cabo Verde. Note that we use the term diglossic in the ideological sense, to refer to ideologies about language hierarchies found in Cabo Verdean society, rather than in line with older academic, sociolinguistic understandings of language and society (e.g. Ferguson, 1959; Fishman, 1967). Almost all of our participants, regardless of their profession or positioning related to the bilingual initiative, made some reference to the way CVL forms an essential part of the Cabo Verdean identity: “I always say that a Cabo Verdean person eats in Kriolu, thinks in Kriolu, sleeps in Kriolu. So [he or she] does everything in Kriolu” (teacher).6

When we began coding our data, we found that ideologies that associate language with identity and questions of educational efficacy were often quite closely intertwined:

[Kriolu] is our language, it’s what we learn, what we think, and the language of our heart! Now imagine that we want to learn more and better, running away from our identity, running away from what is ours. We think first in Cabo Verdean, then later we try to translate for others […] now when a primary-aged child enters the classroom and the teacher starts to speak only in Portuguese, it’s a shock! (bilingual teacher)

For this teacher, who participated in the bilingual pilot project, the monolingual Portuguese-medium school’s position in the historically diglossic language distribution is conceptualised as “running away from” Cabo Verdean identity, but also results in a sort of culture shock for children when they enter school. Like many other participants, he used the terms Cabo Verdean and Kriolu interchangeably to indicate the community language, implicitly establishing a link between the language and the sense of being Cabo Verdean.

Some of the other teachers argued that the “shock” upon finding oneself suddenly immersed in a language different from that used in the community, and which surely is not conducive to learning, is not the same for all children. The other bilingual project teacher we interviewed criticised the school’s participation in a broader hegemonic sociolinguistic structure in which some (but not all) children are inundated with Portuguese-language television programmes that do not reflect their own experiences—“they bring us programmes, cartoons, so there the children have contact with the Portuguese language. But beyond that, maybe just those that belong to the elite social classes.” In this case we have a double-bind situation: children from families with greater socioeconomic means are more likely to access media consistent with the school’s language, but at the cost of moving away from their own experience. Children from economically disadvantaged families avoid this early cultural conflict, but eventually suffer more from the initial encounter with schooling in what for them is a more completely foreign language medium.

The activist we interviewed drew upon his own experience as a child to draw a similar relationship between language, identity, and school success: “The same child who considers himself to be intelligent until the moment he enters the school, becomes an idiot, because he can’t say anything”. In these scenarios, language is considered a resource for understanding and participating in school practice and is strongly linked to issues of self-esteem. As one participant pointed out, the mere act of discovering that CVL is a rule-governed language just like Portuguese had significant implications for her sense of self, as a speaker of the minoritised language:

Learning to write and learning the rules of Kriolu raises our self-esteem. I ended up liking it. Initially I found it a bit, well … but afterwards I started to like it, and knowing that our language has rules. It can be learned orally, as well as by writing. I felt puffed up, in the sense of important. (teacher)

Such comments illustrate the degree to which a deficit ideology concerning CVL is entrenched in Cabo Verdean society, to the extent that these educated professionals recognise that they had implicitly shared it. These teachers’ discourses seem to echo the access paradox, while at the same time drawing upon other kinds of language ideology: they reinforce the relegation of languages to separate and hierarchical domains (home/school) and the unifying notion of a national language, while at the same time drawing upon broader academic and policy arguments in favour of bilingual and L1-based education (Lawrence Gordon & Harvey, 2018; Ouane & Glanz, 2010; UNESCO, 2016). In fact, both teachers who had participated in the bilingual pilot project, both politicians (regardless of party affiliation), the language activist, and many of the other teachers all made specific reference to the documented linguistic and educational advantages demonstrated by the children in the bilingual programme, which included their performance on a written test of Portuguese. These narratives reveal the strength of language ideologies and their inertia even when challenged by scientific evidence.

Standard language ideology—whereby languages are viewed as discrete and finalised entities with distinct linguistic codes that should not be mixed—also emerged frequently in our data. Several teachers argued that studying the structure and written form of the community language allowed them to grasp more clearly the formal register of the colonial language, therefore avoiding interference between two language varieties that are more closely related in some aspects than others: “Certain grammatical errors, even misspellings, that students make when they are writing [in Portuguese] are actually due to a lack of knowledge of the Cabo Verdean language” (bilingual teacher). In the discourse of both bilingual teachers and many of the other participants, this particular language ideology is firmly entrenched in a broader association between language and national identity—these educators fear that abandoning one’s L1 at the cost of schooling in the colonial will language result, ironically, in failing to properly learn either:

The students enter the school with a language. It exists, this language exists, they have learned some things since they were born, since they began to speak, and now we will erase it?! Erase that and introduce Portuguese?! Then there are serious difficulties in the Portuguese language because we are forgetting the Cabo Verdean language. The introduction of the Cabo Verdean language in the first year of school is fundamental to improve Portuguese. (bilingual teacher)

[The bilingual pilot project] encouraged the learning of the Portuguese language based on an exhaustive exploration of what the Cabo Verdean language is: the language in which students think, interact, create, dream, and delight – learning about Cabo Verdean reality. (former education official)

One of the bilingual teachers described a presentation he had made about the bilingual programme’s results, where children’s linguistic competence in written CVL and Portuguese produced a visible emotional impact on some audience members: “I presented a text of a first-year student written in the Cabo Verdean language, also translated into Portuguese, and some people […] ended up shedding some tears”. For this teacher, and for the conference delegates he described, bilingual competence is not a strictly academic matter, but resonates with a long history of oppression, so that linguistic and educational success is tightly interwoven with personal and historical associations with these languages.

The Access Paradox and a Discourse of Postcolonial Resignation

At the same time, some of our respondents also drew upon discourses that served to support the postcolonial monolingual status quo (Ouane & Glanz, 2010). We will focus in this section on data from the two group interviews, in order to analyse specifically the discourse of teachers who did not participate in the bilingual programme, and therefore were not stakeholders in the same way as participating teachers, politicians, and language activists. These teachers are outsiders, but with a professional positioning that renders them particularly experienced and interested in the issue of bilingual education.

Compared with those described above, anti-bilingual education discourses were generally weaker in the interviews, reduced in terms of presence as well as intensity, and were often related in the third person, as beliefs held by particular others or collectively as a popular belief that they did not necessarily share. When deployed in the first person, they tended to lack the emotional intensity evident in those defending the “language of our hearts”. These declarations usually took on a more calculated, strategic tone. For example, while many respondents felt that CVL should be used in schooling because it was the language of the community and therefore would afford home-school continuity, one respondent provided exactly the opposite argument for the same reasons. He also recognised that Portuguese was scarcely present in the children’s home lives, and therefore argued that it should be emphasised in schooling in order to compensate for what he perceived as an imbalance favouring CVL:

We have the orientation to insist very much on teaching Portuguese because it is an opportunity. Because we have few hours. During practically twenty-four hours the students are practising Kriolu, and in four hours of class it’s best to strive to work in Portuguese. (teacher)

This teacher draws upon discourses of balance and compensation reminiscent of the US English-only movement, which argued that children of migrant backgrounds needed to be immersed in English-medium schooling order to compensate for their lack of exposure in the home. The arguments from the US context went so far as to compare bilingual education with child abuse, as it was seen to deprive children of the social benefits afforded by the socially dominant language (May, 2014). In our research, the discourse of compensation (in favour of Portuguese) never approached the level of anger and anxiety that characterised the US debates; even the teacher who spoke in favour of teaching (mostly) in Portuguese also recognised the culture shock experienced by CVL-speaking children when they encounter Portuguese-medium schooling for the first time: “Because [of ] the students […] above all the first-year students who do not yet know Portuguese, it is very difficult […] to teach a class without resorting to Kriolu”. At the same time, his word choice implies a value difference: while Portuguese is a language of opportunity, to be strived for, Kriolu is a language of necessity (to be resorted to) temporarily, until children can understand Portuguese.

These kinds of arguments reflect the recognition of a paradox of access: Portuguese is the language of opportunity that must be fostered if children are to thrive, especially in the economic sense of achievement. There is a conviction that a choice is forced on Cabo Verdean society, a choice that research on bilingual education suggests is a false one (Cummins, 2005). Nevertheless, this discourse reflects the popular logic that language learning is subtractive: learning one language detracts from, rather than facilitates, learning another.

One teacher described how issues of bilingual education were currently being debated in popular circles, specifically citing a Facebook forum that she had been following. She told us said that the demonstrated academic achievement of the bilingual students was one argument used in favour of continuing the project, while others argued against bilingual education on the basis that children have serious deficiencies in Portuguese and so need to learn this language first, and then CVL. These posters felt that simultaneously learning in both languages might inhibit the learning of the most important language (Portuguese), a commonly held understanding of bilingual education that is shared by teachers and students.

We also identified a thread of tacit acceptance of the status quo in many teachers’ declarations, even when they expressed support for bilingual education involving the children’s home language. One teacher, for example, argued that they are so strongly conditioned to teach in Portuguese, only “resorting to” CVL in cases of necessity, that it would be difficult to change:

It is that we have not, for the most part we’ve been trained that the teaching of the Portuguese language has to be in the classroom. So usually when this happens, if the students are speaking Kriolu, we usually impede this [behaviour]. We say, better try to speak Portuguese, and we make the correction. We never motivate the students to speak Kriolu – this creates a certain tension.

This kind of argument reflects a sense of language hegemony supported by habit (the way we’ve always done it), a system inherently resistant to change. This same teacher also spoke of a tacit acceptance of Portuguese as a higher register, appropriate for use in institutional spaces:

And there is another question as well: usually in any institution that we are in, normally the language spoken is the official language, it is the Portuguese language […] Therefore, we have practically devalued our mother tongue […] That our mother tongue is in disuse, the fact that it is not used in different official places will result in it eventually losing value in relation to other languages. With respect to Portuguese.

He expressed concern at what he saw as the “devaluing” of his “mother-tongue”, and at the same time implicitly included himself in these processes through his use of the first-person plural (“we”). Another teacher described Portuguese-medium instruction as the only reasonable response to what he sees as an implacable sociolinguistic reality:

It is the language of communication, it is the language that the student uses in books, in official communication […] Fortunately or unfortunately we have [Portuguese] as our official language. So they have to learn it […] I recommend that the teacher should teach more in Portuguese.

At the same time, the pure pragmatism of this recommendation is slightly conditioned by the use of the phrase “Fortunately or unfortunately”: this teacher refuses to offer his own value judgement, and in so doing conveys a sense of resignation. In the individual and group interviews, when asked whether the teachers’ union had advocated for teaching in CVL, there was a clear consensus that bilingual education was not an issue on the agenda. One teacher explained that teachers’ collective action was more focused on their teaching conditions and salaries: “Usually the union and these things are more concerned with those issues that have to do with pay raises and teachers’ rights.”

The Ambiguity and Fragility of Political Discourses

The MpD education official, himself a former teacher, echoed many of the ideologies about the nature of language(s) and their relationship with national identity which we have described above, and which were pervasive throughout the interviews: he saw CVL as an important part of his heritage and described CVL language teaching in terms of cultural maintenance: “I have to start by telling you that for me language is … language is one of the elements of the culture that I identify with. Trying to teach my language means that I am perpetuating what is mine”. The official invoked the notion of symbolic violence to describe Cabo Verde’s monolingual (post)colonial history. At the same time, he used the same concept to describe efforts to redress this historic diglossia. Although this is not the case with the bilingual programme, which uses both languages, he referred to teaching entirely in CVL as symbolic violence: “Administering everything in Kriolu is more symbolic violence”. At the same time, he spoke of the positive results of the bilingual programme, highlighting the improved academic achievement of students in the bilingual programme compared to students in the standard monolingual Portuguese programme: “I made a comparison … the result is … is …. No … the class where [bilingual project teacher] worked with them … comparing the results is far better … far better”.

As with the teachers’ discourse, we see national identity and academic success intertwined. The argument here clearly reflects that of the access paradox: it would amount to symbolic violence to impose exclusively either of the two languages. The official pointed out the superior Portuguese-language achievement that he had witnessed first-hand among the bilingual students (which was confirmed by the two participating teachers and some others), and at the same time denounced the programme for committing the symbolic violence of denying Portuguese proficiency. Furthermore, he characterised the bilingual project as one of CVL immersion, which is clearly discredited by the results in Portuguese proficiency. Although the programme’s bilingual nature was designed to avoid the paradox, it is discursively reconstructed to position national identity (learning the national language associated with independence from the former colony and local identity) at odds with academic success (learning the postcolonial language and thereby developing an instrumentally valued skill that may lead to economic advancement).

The government official, in line with some of the teaching staff we spoke with, positioned CVL as a “resource for learning” in the classroom that could form part of a “curricular enrichment programme”, where Portuguese is maintained as the vehicular language of the classroom, and the children’s home language is “resorted to” as needed to foster comprehension. He provided as an example the practice he supported among his trainees as a supervising teacher:

My students used the Kriolu language as a resource language […] What matters is that inside the classroom there must be an environment for the construction of knowledge, whether it is with one language or another. I understand that it is the teacher’s responsibility. Why? Because I as a teacher worry about the results, the result is that … it would be good if all my students were at the top of the line.

There was a reluctance regarding the use of written CVL as he noted that “when using the Cabo Verdean language as a resource in a written way, it is even more difficult for students to learn”.

Here we see an ideological divergence with respect to the teachers who participated in the bilingual programme: they coincide in the more pedagogical argument that using the children’s home language provides comprehension, and therefore supports academic success. Yet the official implicitly embraces the language hierarchy challenged by these teachers, who equate incorporating CVL as a language of instruction with language equity. Here the government official appears to echo ideologies underpinning early-exit bilingual programmes. Along with some of the other teachers we interviewed, he described the community language as a pedagogical resource, but did not situate academic fluency and raising the status of the community language among curricular objectives. The access paradox is resolved in a way that prioritises access to the colonial (global, European) language. There is also a monolingual ideology at play here, which may be conditioned by economic concerns: the possibility of becoming fluently bilingual, one of the principal ideologies underpinning late-exit or development bilingual programmes, is not contemplated.

In his interview, the education official also noted that government support for the programme had not been withdrawn, and he spoke enthusiastically of its eventual resumption:

I believe that in the future there will be a return, it will happen for sure! Because right now it is in a phase of study between … in a phase of study … both the Ministry of Education and the Ministry of Culture are looking into what best suits our country.

He argued, however, that in order for the bilingual programme to continue, it must be stopped long enough to allow for teacher training, development of materials, and the establishment of a universal standard form of CVL to be used in all classrooms. While there is currently a standard alphabet (ALUPEC), this phonetic system does not dictate universal spelling rules, and there is dialectical variation among the different island groups that comprise the nation. While dialectical variation is a natural aspect of any language, this linguistic reality does conflict with more purist ideologies that insist on a unified standard. He insisted on the necessity of establishing such a universal classroom standard, but also reflected on the challenges associated with language standardisation:

If the standard is determined along with the Ministry of Culture, which one is it? Because also people from [a particular area] can say, “No! the standard should be our language”, from [another area] you can say, “No”, [another area] the same thing, [another area] the same thing. So we get several variants.

What emerges here is a double-bind situation: if the project must remain on hold until a consensus is reached on a standard form of CVL, then it is hard to envisage when the project will resume. This purist ideology has circulated in popular discourse, as described by our other interviewees:

The only concern is perhaps that we have several dialects in Cabo Verde, because [one area] speaks Kriolu in one way. (teacher)

First we have to have a dictionary in Kriolu […] there has to be a standardisation of the Kriolu language at the national level. (teacher)

In these examples, the standard language ideology seems to have been taken up uncritically as an argument for delaying bilingual education in Cabo Verde.

Discussion and Conclusions

Multilingual education and language policies are key to achieving democracy and development and avoiding decision-making being reserved for elite members of a society (Alexander, 2007). As Champion (1974, p. 4) notes, as long as the language of instruction is the language of the former colonial power, “the school remains colonial and abstract and whatever innovations are introduced remain superficial”. Monolingual education serves to entrench the disconnect between school and society, impacting on effective learning and progress. Cabo Verde is frequently cited in the literature as one of the few “monolingual” African countries—along with Lesotho and Swaziland—but virtually the entire population speaks CVL. It is striking that there continues to be a rejection of CVL in formal and academic domains.

Using this specific case to explore the access paradox, our analysis of discourses produced regarding the introduction of bilingual education illustrates how multiple intersecting ideologies about language and learning may complicate the resolution of this paradox. Some of the main ideologies that would support bilingual education are clearly expressed by all the participants in our study. There was a clear consensus that CVL is associated with Cabo Verdean national identity, and a critical consciousness of historical hegemony that has relegated the community language to the bottom of the language hierarchy. While the access paradox does not speak to pedagogical concerns, we also see ideologies that support bilingual education in terms of the perceived psychological and educational damage of language submersion. These were often closely connected to language and identity arguments, so that academic and identity-building projects were seen as inseparable goals for bilingual education.

At the same time, competing ideologies serve to weaken efforts to introduce bilingual education in Cabo Verdean schools. While bilingual education seeks to resolve the access paradox, so that children are not forced to choose between their community and global languages, a monolingual ideology was present, particularly in the discourses from the political sector. Finally, the ideology of language standardisation figured strongly in political discourse and was also taken up uncritically in teachers’ discourse and may well prove the most effective in blocking the future development of bilingual education in Cabo Verde. Other, more logistical concerns might also hinder progress, and these include teachers’ own monolingual (Portuguese) education. None of them reported having any training in bilingual education, and most specifically cited their own lack of experience with formal registers of CVL (oral and written) as a problem.

Notwithstanding the focus on subtractive, submersive, monolingual education in Cabo Verde and other postcolonial African countries, the teachers we interviewed revealed that multilingual practices take place in classrooms every day, albeit without official support. Local languages and code-switching practices are inevitably used in the classroom to ensure that effective learning and two-way interactions can take place (Heugh, 2021). These practices, however, are deemed illegitimate and not suitable for formal education (Heugh, 2021). Code-switching has been associated with characteristics such as laziness and a rejection of the authority of the state, while monolingualism and linguistic purity are positioned as attributes of a loyal citizen (Stroud, 2007). As a result of the stigmatisation of hybrid linguistic practices, there is usually great emphasis placed on the separation of languages within the education system. Probyn (2009) reflects on covert code-switching practices that take place in multilingual classrooms and emphasises how teachers often must resort to “smuggling the vernacular into the classroom”. Our study, then, corroborates previous studies which found that in most African classrooms, some form of bilingual education is already taking place (Ouane & Glanz, 2010); teachers use the local language(s) to ensure effective communication and make sure that students understand what is being taught (Ouane & Glanz, 2010), but this is usually limited to oral communication.

In Cabo Verde as well as other postcolonial multilingual contexts, the informal and relatively weak incorporation of the minoritised language in schooling contexts could relatively easily be transformed into the kind of rigorous and systematic bilingual education supported by linguistics and educational research. Based on the results of our study, we identify teacher training for bilingual education as an area for further exploration. In the Cabo Verdean and similar postcolonial contexts, this might focus on three areas: linguistic competency (especially written), bilingual education (philosophy, programmes, and research results in other contexts, techniques), and the specific local sociolinguistic context. This training may be made available for all teachers, but especially for those participating in the bilingual project. It might take the form of in-service training or be incorporated into the preservice teacher certification process. In the Cabo Verdean context, what is currently missing is policy building based on academic research, which would include a consistent and formal implementation of bilingual education across the country, teacher support and training, and a clear discussion and negotiation of this model with families and local communities.