Designing for Complexity in Mother Tongue Or First Language (L1)-Based Multilingual Education Programs

Designing for Complexity 41

Designing for Complexity in Mother Tongue or First Language (L1)-Based Multilingual Education Programs

Corrie Blankenbeckler Creative Associates International

Abstract Mother-tongue or first language (L1)-based multilingual education programs are necessarily complex and may require a more nonlinear approach to program design. These programs operate within and act upon a range of psycholinguistic, sociolinguistic, and sociopolitical issues that include language structure and literacy assessment, language policy and politics, and cultural and social behavior change linked to literacy expansion. The broad use of one-size-fits-all outcomes-based design approaches for L1-based multilingual education programs often result in designs that are retrofitted to new country settings and ill-suited to the context in which they are implemented. This paper looks at some of the many features that can be used to inform the development of L1-based multilingual education in the context of early literacy programming. Specifically, it examines the use of alternative approaches in the development of flexible theory of change design that integrate early literacy and L1-based multilingual education program design frameworks to more suitably address the sociolinguistic, sociopolitical, and psycholinguistic assumptions underpinning multilingual education approaches.

Keywords Mother-tongue; Multilingual education; Complexity; Nonlinear design; Early literacy; Theory of change; Sociolinguistics; Psycholinguistics; Language policy and planning

Introduction significantly depending on the funding Life is not simple, but many of the logic models used mechanisms and context. Glouberman and in programme theory evaluation are. Is it a problem Zimmerman (2002) draw distinctions between to represent reality as a simple causal model of boxes simple, complicated, and complex program and arrows, or should the logic models we use address the complexity of life—and if so, how? theories that are helpful in determining (Rogers, 2008, p. 29) interventions that may require evolving design as effects or outcomes emerge through the Program theory of change is intended to course of implementation. To use their guide purposeful program planning and examples, a simple theory can be compared to a evaluation. Variously described as theory of recipe, easy to replicate, provided you have the change, theory-based evaluation, logic model, right tools and ingredients. Sending a rocket to and many other variations, program theories the moon is far more complicated to design and outline causal linkages between different aspects difficult to evaluate given the number of of the program (inputs, activities, outputs and variables involved. Raising a child brings in outcomes) used to evaluate program levels of complexity that preempt common effectiveness or impact (Rogers, 2008; Weiss, agreement on a predetermined design for 1998). A theory of change is commonly used as a success. In an analysis of how to address the heuristic device, an abstract framework, or a complicated and the complex in program design, model that informs thinking and predictions Rogers (2008) underscores the need for models about social change. The models used in the that reflect the organic complexity that development of theories of change can vary underpins human interaction. Programs

______Global Education Review is a publication of The School of Education at Mercy College, New York. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License, permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. Citation: Blankenbeckler, Corrie (2020). Designing for Complexity in Mother Tongue or First Language (L1)-Based Multilingual Education Programs. Global Education Review, 7 (1). 41-VVVVV 42 Global Education Review 7 (1)

designed to address social behavior change can [Kenya]; Skutnabb-Kangas & Heugh, 2012, tend to oversimplify with a single linear design [Ethiopia]). The opportunity to learn in the that ignore related parallel causes, or may mother tongue or first language (L1) is require more iterative, nonlinear approaches recognized as a right (United Nations, 2008) that open opportunity for unexpected learning and a benefit (UNESCO, 2003), though the and adaptability. The language contexts in which matter is not without continued dialogue and mother tongue or first language (L1)-based debate (May, 2005; Skutnabb-Kangas & Heugh, multilingual education programs operate 2012; Wiley, 2012). Although many language arguably fall into more complicated or complex policies around the world have been amended to categories demanding nonlinear approaches to accommodate the linguistic diversity that is program theory design. reflective of a language environment, over half of This paper examines the complexities of the the world’s children continue to be taught in a language environments in which L1-based language that is different from what they speak multilingual education programs operate that at home (Ouane & Glantz, 2010). are often overlooked in programs that focus on Hornberger (2009) observes that early literacy, with out-of-the-box outcomes- multilingualism, not monolingualism, has been based design focused on quick wins, most often the natural state of human societies throughout manifest in measurable improvements in history and that the imposition of a single reading subtask scores (i.e., letter recognition) in language serves as a cultural limitation. It is true target schools (Bartlett et al., 2015; Benson, that despite the demand for, and increasing 2004). The restricted focus of early literacy dominance of, English and other global intervention programs on mother-tongue languages (Clayton, 2008; Hornberger, 2009; outcomes in the early grades, while showing Ricento, 2000), linguistic diversity remains a some improvements in the short term, can defining characteristic of the global context. discount larger psycholinguistic, sociopolitical, Although language statistics often obscure the and sociolinguistic issues that are present in complexity of linguistic heritage or language multilingual environments. A limited focus can type and use, the numbers are still staggering lead to an overall, longer term negative impact (Skutnabb-Kangas & Heugh, 2012, p. 12). There on student learning, especially in contexts where are between 1,500 and 2,000 African languages, there is an early transition from the first over 80 estimated in Ghana alone (Mann & language to the second or national language Dalby, 2017). India has 447 living languages, 64 (Alidou et al., 2006; Baker, 2011; Hornberger, of which are institutional (Eberhard et al., 2019). 2002). In the United Kingdom, an estimated one in six children speak a language other than English at L1-Based Multilingual Early Literacy home (Tinsley, 2013). Given these facts, it The benefits of learning in a familiar should not be surprising that an estimated 40% language have long been recognized (UNESCO, of the world’s people lack access to education in 1953), and there is a large body of evidence from a language they can speak and understand diverse contexts that has demonstrated the (UNESCO, 2016). Of the 6,000 -7,000 languages positive effects of learning in the mother tongue spoken around the world, only a few hundred in diverse contexts (e.g., Alidou & Brock-Utne, are used in education (Walter & Benson, 2012). 2011, [Niger]; Benson, 2004b, [Mozambique]; There is general consensus that the use of Hovens, 2002, [West Africa]; Piper et al., 2016a, the mother tongue as medium of instruction for Designing for Complexity 43

foundational literacy has a tremendously and numeracy. In other cases, the language of positive impact on education quality and instruction is the national or official language academic growth of individual learners, (L2) and the mother tongue is used by the especially in low-income contexts (Alidou et al., teacher only to ease learning in the L2. In 2006; Benson, 2004a; Cummins, 2000; multilingual contexts where there are local Skutnabb-Kangas, 2000). While monolingual languages spoken that differ from the national education is still standard practice around the language, it is rare that the L1 is used as world (Benson, 2014), a number of early grade language of instruction beyond primary school, literacy programs now focus on the acquisition and it is very often the case that it is used only in of reading skills in the mother tongue prior to the early grades—grades one, two, and transitioning to a second language, usually a sometimes three (Alidou et al., 2006). national language (e.g., Amharic in Ethiopia) or The arguments for the use of L1 in early global language (e.g., English, French, or grades is tied to strong evidence showing that Portuguese). Definitions of “mother tongue” in the L1 facilitates the acquisition of reading skills these programs can vary depending on context. in both the L1 and L2 (Bialystok, 2011; Collier & According to Skutnabb-Kangas (2000), the Thomas, 2004; Cummins, 1984;). Cummins’ mother tongue can refer to a language that a (1984) linguistic interdependence hypothesis child may have learned first or knows best, but explains a process of transfer of L1 language and can also be the language the child most identifies literacy knowledge into the L2, whereby L2 with (or is identified with by others) or uses the proficiency is heavily dependent on the most. In education, the L1 is generally defined as capabilities children have already developed in the language that a child understands and the L1. Learning to read first in an L1 does not speaks well enough to be able to learn put young children at a disadvantage; in fact, the successfully at grade-level (Benson & Kosonen, learning transfer occurs both ways between L1 2012). In this case, the L1 is not always the and L2. The transfer should not be rushed and language that the child speaks at home, learned must pay proper attention to oral language first, or knows best. In many contexts, the development in both languages. Foundational selected language of instruction (referred to as literacy skills in the early grades of primary the L1), may be a language from the community school must be maintained and developed that is well-understood by the child, but is not in throughout early primary education in order to fact the child’s mother tongue. In this paper, I succeed in school (Benson, 2004b; Stanovich, use the definition provided by UNESCO (2003): 1986). L1 as primary or first language. L2 will be used The task of learning to read (in any to describe a second language introduced in language) is not easy or straightforward, education, and Lx may be used to describe requiring a complex set of processes working multilingual contexts where a third or additional together (visual, linguistic, cognitive) in order language is used. for the brain to make sense of the markings that L1-based MLE describes programs that form letters and words (Dehaene, 2009; Wolf, prioritize learning in the L1 in a multilingual 2008). Education systems struggle to provide environment (Bender et al., 2005; UNESCO, learners with the tools they need to acquire these 2003). In some contexts, the L1 is introduced as skills, and many learners quickly fall behind, the language of instruction in the early grades to especially in low-resource contexts, where facilitate learning the fundamentals of literacy learner-appropriate printed resources are a 44 Global Education Review 7 (1)

scarce commodity. The negative effects of practical definition has been broadly used to delayed reading skills compound later academic inform program development and pedagogical challenges, often leading to high dropout rates, approaches. particularly among poor and disadvantaged The reductionistic way in which these groups (Stanovich, 1986). In a call to arms that approaches are applied in diverse contexts have laid some of the groundwork for the Sustainable been criticized (Hoffman, 2009). In Development Goals, UNESCO (2014) declared a international program practice, the simple view “global learning crisis”: 250 million children are is at times erroneously used to denote a not learning to read and write, including the half relationship between fluency (defined and that have attended early primary grades. measured as automaticity) and reading Progress is slow, and nearly 50% of children and comprehension, rather than decoding and youth are still not achieving minimum linguistic comprehension for reading proficiency levels in reading and mathematics comprehension. Scholars have attributed such (UNESCO, 2019). confusion to the widespread use of the Early There are two major challenges that have Grade Reading Assessment (EGRA) developed emerged over the past few decades: 1) an by the Research Triangle Institute in 2005 unstated reliance on monolingual, English (Benson, 2013; Dowd & Bartlett, 2019; Graham language-based early literacy development and & van Ginkle, 2014). The EGRA is based on an assessment research to guide international adapted version of Dynamic Indicators of Basic program design, and 2) heavy emphasis on Early Literacy Skills (DIBELS) designed for development of a subset of foundational literacy English literacy learners in the United States. skills (e.g., phonological awareness and EGRA was initially developed as a diagnostic phonics), often at the expense of critical focus on instrument that has been used to shine a more wholistic development of a child’s spotlight on the extent of poor reading skills in receptive and communicative language low-income contexts (Benson &Wong, 2015; development in both the L1 and L2. Many early Dubeck & Gove, 2015; Gove & Cvelich, 2011). As grade reading programs introduced in a diagnostic tool, the results are used to inform international contexts over the last two decades curricula and training reform efforts, and define are guided by the principles drawn from the L1-based MLE intervention design (Pflepsen, findings of the National Reading Panel (2000), 2015). Beyond diagnostics, there is now which sets out five core component skills of widespread reliance on the EGRA as a reading: phonemic awareness, phonics, framework by which to structure and measure vocabulary knowledge, oral reading fluency, and progress in early grade literacy program comprehension. While building on one another, implementation in a wide range of multilingual each of these skills require a discrete set of contexts, which has its limitations. approaches to teach effectively, requiring Substantial floor effects on the fluency explicit instructional practices (Archer & benchmark in EGRA results have led to Hughes, 2011). In Hoover and Gough’s (1990) interventions that target what Bartlett, Dowd, “simple view of reading,” the complex process of and Jonason (2015) categorize as emergent and learning to read is distilled into two categories: decoding skills, stressing “phonics and decoding and linguistic comprehension. Each phonemic awareness more than comprehension” category carries with it a complex set of (p. 309). Scholars have also pointed out that the processes needed to acquire literacy. This EGRA assessment does not assess prosody, and Designing for Complexity 45

fluency rates (correct words per minute) have These studies are contributing to a broader been used as a proxy for reading comprehension, base of evidence that demands a more nuanced an assumption that has its limitations when psycholinguistic and context-driven approach to comparing fluency across languages with address literacy acquisition in multilingual variable word length and grapheme complexity contexts that have not yet been adequately (Dowd & Bartlett, 2019). Graham and van addressed by the EGRA and other global Ginkle (2014) provide a thorough critique of standards. Despite the growing body of research, these limitations in non-English language study of biliteracy and multi-literacy contexts, noting that the use of correct words per development is still very much in the nascent minute as a benchmark fails to account for the stages. Successful biliteracy or multiliteracy “range of differences between languages and development varies by context, language, and orthographies that may influence reading policy, but is also dependent on language- acquisition” (p. 245). Length of the words specific pedagogical approaches to literacy (agglutination), syllable structure, linguistic acquisition. More research is needed to provide transparency, and tone all play a role in the time language- and context-specific definition of the needed to read and understand a passage, and it threshold and independence hypotheses that is difficult to compare different languages undergird cross-linguistic transfer and biliteracy against a common, timed standard. acquisition among non-dominant languages and Attention to the structure of the languages marginalized populations. Greater consideration plays an important role in biliteracy for language policy and planning will be needed development. Linguistic distance and writing to address and sustain mother tongue-based systems can present substantial challenges for L1 programs that are able to deliver biliteracy and to L2 transfer (Koda, 2007; Perfetti & Dunlap, multiliteracy outcomes for successful learning 2008; Tan et al., 2003). Researchers have beyond the early grades. examined the cross-linguistic differences in phonological awareness and receptive Language Policy and Planning vocabulary as predictors of word and letter The politics of language deeply influence the reading tasks and found that the range of success of multilingual education initiatives. difference is dependent as much on the structure Languages develop from and are situated within of the language as on relative status of the specific communities and cultures, and can be language—L1 or L2 as medium of instruction understood as the code to understanding a (Jasińska et al., 2019; Ziegler & Goswami, particular culture or society (Wardhaugh, 2011). 2006). Studies have also shown that students Language acts or policies have a profound need to develop a sufficient oral language impact on the social construct of the affected threshold in the L1 in order to transfer the communities and cultures. In the African reading skills to the L2 (Hovens, 2002; Koda & context, Ouane and Glantz (2010) have Zehler, 2008; Piper et al., 2016a). Recent associated the investment in African languages research on the transfer of L1 to L2 literacy skills and multilingual education with the “deep social in India has demonstrated that a higher transformation induced by a political, cultural threshold is needed for L2 oral language and development project and an education development in order to optimize the transfer of reform agenda” (p. 48). Language policies are L1 literacy skills for reading in the L2 often influenced by local language rights, (Nakamura et al., 2018). linguistic development of the relevant languages, 46 Global Education Review 7 (1)

or other ideological factors. The relationship of in Canhanga & Banda, 2017), purification language to identity and culture also carries with (removal of borrowings, rise of language it the power to marginalize and oppress academies), and terminology development (new communities, which has been done time and term, old term with new meaning, borrowed again throughout the history of human term, calque). Acquisition planning is concerned development (Skutnabb-Kangas, 2000). with language distribution and use, such as Educational initiatives that aim toward medium of instruction in education (Tollefson, multilingualism have the potential to empower 2008). L1-based MLE programs commonly fall those who have been oppressed through under acquisition planning, seeing language as a language policies, and they also hold the resource for learning (see the discussion on potential to activate local voices, but failure to orientation in language planning in Ruiz, 1984). address language policy and planning can result L1-based MLE programs will inevitably in an experimental approach that does not have interact—by design or accident—with status and lasting impact, or may have unintended negative corpus planning, especially when non-dominant consequences. Unfortunately, many programs languages are introduced as the medium of fail to address the policy and planning required instruction. The official status of a language— for meaningful and sustained implementation of language policy—does not automatically new languages into the school system (Ansah, translate into a language of instruction (LOI) 2014; Hornberger, 2006a, 2009; Trudell & policy. In Morocco, Berber (Tamazigt) was Piper, 2013). regarded as a minority local language until its Language Policy and Planning (LPP) has official recognition in 2011 (Maddy-Weitzman, been variously employed over time by scholars 2011). However, despite efforts to introduce the and practitioners to describe and inform Berber language into the public-school system, behavior change with regard to language the languages of instruction are still Standard management (Ricento, 2000; Ruiz, 1984; Arabic and French, while Berber is taught as a Tollefson, 2002). LPP is commonly formulated subject only in a certain select number of schools around three overlapping and interdependent (Johnson, 2013). In other cases, minority areas: status planning, corpus planning, and languages are not officially recognized, which acquisition planning. Each of these has taken on leaves many children without access to various forms and approaches depending on the education in their mother tongue. There are context and use (Hornberger, 2002). Status between 50 and 80 indigenous languages spoken planning, also described as the function of in Ghana, although only eleven are recognized language in society, deals with the politics of by the government, and English remains the linguistic imperialism, language revitalization, official language of state institutions (Ansah, language use in official domain, and issues of 2014; Eberhard et al., 2019). For sustainability language selection and choice (Clayton, 2008; and quality of programming, L1-based MLE Hornberger, 2006b; Skutnabb-Kangas, 2000). programs must enter into policy level In corpus planning, efforts are generally directed discussions in order to advocate for the models toward specific areas of linguistic structure and of education, all of which will depend on the variants, including standardization, theoretical or ideological orientations to which graphization, choice of script, and orthographic the program ascribes. conventions (see for example the discussion of Language of instruction policies also vary in orthographic standards among Bantu languages purpose and scope, and rarely do these policies Designing for Complexity 47

address issues of language-specific approaches weaken a program before it has even begun to literacy instruction. Baker (2011) presents a (Piper, Zuilkowski & Ong’ele, 2016). In short, comprehensive typology of the multilingual L1-based MLE programs can inadvertently education models from fully L1 monolingual problematize languages through exclusion (submersion or segregationist), to weak (through ignorance or by design), or by failing to transitional L1 to L2 models (usually early exit), address some of these policy issues early on. to strong additive (immersion, maintenance two-way, dual medium, and mainstream) L1-based MLE Programming in Additive approaches are considered most Context effective, integrating local languages alongside L1-based MLE programs have expanded the L2 for as long as possible, even through rapidly over the past two decades. As of 2016, secondary education (Alidou et al, 2006; Benson with substantial support from the United States & Kosonen, 2012). By contrast, the more Agency for International Development (USAID), common early exit model assumes that students as well as the World Bank and others, the EGRA will achieve mastery of the L2 at a very young has been implemented in 72 countries and an age (primary grade 4), even though they may estimated 129 languages (Dowd & Barlett, 2019). have little experience of the language outside of The emphasis of these programs has been L1- the school context (Alidou, et al., 2006; based early grade literacy instruction. Programs Cummins, 1979). In addition to the above have highlighted the paramount importance of challenges, many curricula attempt to teach the use of the mother tongue in early literacy students to read and write in both the L1 and L2 acquisition and have seen some success in simultaneously, as is the policy in Bangladesh moving from pilot to policy-driven scalable (Hamid & Erling, 2016). models (Gove & Wetterberg, 2011). The USAID Whatever the model, the many other 2011 Education Strategy made a strong case for practical issues of implementation of L1-based L1 learning and the need for USAID MLE programs will also interact with other LPP programming designed in accordance with domains. Teacher selection and retention (e.g., “appropriate language policies,” and “where mobility issues) will raise questions of the these policies do not exist, USAID should engage viability of L1-based MLE programs in the short- in policy dialogue with host country government and long-term (e.g., teacher language capacity and partners in an attempt to improve policy, as and willingness to teach in the local language, on other technical issues” (USAID, 2012, p. 4). language policy change and geographic shifts of In 2015, the USAID-funded Global Reading language of instruction policies; Heugh, 2008). Network published a document designed in Curriculum and textbook development can very “practical response to requests from USAID’s quickly become entangled in orthographic and Africa Missions,” to assist in program design for linguistic variant debates when indigenous L1-based MLE (Pflepsen, 2015, p. 1). The languages are newly standardized, or still in the publication highlights many of the above- process of standardization, with little experience mentioned LPP considerations and provides an or widespread use of the new standardized important list of concrete steps that might be versions (Alidou et al., 2006). Inappropriate or followed in the implementation of USAID- top-down language selection and language funded L1-based MLE programs, including mismatch (spoken language different from transfer of skill across languages, language medium of instruction) can compromise or context and mapping, orthography development 48 Global Education Review 7 (1)

and standardization, teachers’ language numeracy, e.g., Ghana). Improvements in early proficiency, and so on (for the full list, see pp. grade reading may be accompanied by a qualifier 42-43). The USAID 2018 Basic Education such as proficiency or achievement, but without Strategy continues support for L1-based MLE reference to medium of instruction (e.g., in but does not reference the report, nor does it Mozambique, the desired improvement is to provide explicit guidance on how to approach read fluently and with comprehension). The L1-based MLE program design. intermediate results of the program focus on Scholars have used an ecology perspective as materials and instruction (combined or a way to understand and analyze the ideologies separate), systems strengthening and capacity undergirding language policy and practice building (combined or separate), and (Hornberger, 2002, 2006b) and the interaction community engagement and support for between languages “inside and across polities” education. Sometimes a unique fourth or fifth (Creese et al., 2008). Hornberger (2006b) result is included in the program design (e.g., defines three themes in the ecology of language. access in Zambia, assessment in Morocco, First, languages live and die in relation to other professional development of teachers and languages. Second, languages are situated within government accountability and transparency in a specific context—the language environment— Ghana, and vulnerable students and that encompasses all range of human issues such communities in Ethiopia). The EGRA is the as politics, economics, and culture. Third, expected measure for learning outcomes in all languages are not static; they take breath in cases, except in the Philippines (no assessment human hosts, and thus evolve and change, and mentioned), and in Ethiopia (where there are can be endangered. The use of this schema explicit instructions not to conduct an EGRA “recognizes that planning for any one language assessment). None of these design frameworks in a particular context necessarily entails include explicit reference either to L1-based planning for all languages impinging on that instruction or to biliteracy or multilingual one” (Hornberger, 2006b, p. 280). Conversely, education outcomes. programs that focus only on the use of the L1 fail The above results frameworks would fall to address planning for all languages. under Chen and Rossi’s (1989) definition of “a A general review of recent USAID early sparse simple theory,” that is distinguished from grade reading solicitations in recent years shows a “rich theory” (p. 301). To illustrate, a simple heavy emphasis on L1-based reading instruction theory for early grade reading might be that that is not always made explicit in the theory of learner appropriate curriculum and instruction change or results framework, and does not will improve reading. A rich L1-based MLE provide sufficient consideration of the theory, by comparison, might be proposed as multilingual context. The following patterns follows: learner appropriate curriculum emerge from review of a few illustrative program instruction, when developed with stakeholder designs since 2015 in Egypt, Ethiopia, input for biliteracy in the L1 and L2, will enable Cambodia, Ghana, Morocco, Mozambique, learners to read to learn throughout the Philippines, and Zambia (United States remainder of their learning experience. A more Government, n.d.). The objective or goal of the in-depth read of the full solicitations reveals program is typically to improve early grade many implicit assumptions about program reading (sometimes grade specific, e.g., implementation that may result in emergent or Cambodia; and once with the addition of unintended outcomes (e.g., policy environment Designing for Complexity 49

that does not favor local language instruction). selection and community reception of the In the following pages, I provide a closer look at program. There were several languages the implementation of an early grade reading represented in Zambezia not included in the program in Mozambique to highlight some of program, including one with a relatively large the many dips and turns a program may take population of speakers—Echuwabo. One of the that may deviate from the program design, and perceived challenges in implementing bilingual challenge meaningful evaluation of program education at large scale at the time was the fact effectiveness and impact. that the introduction of bilingual programming The USAID-funded Vamos Ler! (Let’s Read!) in individual schools is premised on community program, a five-year program (2016-2021) choice (Blankenbeckler, 2017). Bilingual implemented by Creative Associates education was indeed planned for gradual International and partners, is designed to expansion, but the 2002 curriculum reform address poor reading outcomes by introducing a policy allowed for communities to choose L1-based approach in support of the planned between monolingual, semi-bilingual, or national expansion of bilingual education (Potter bilingual modes in the early grades (MINEDH, & Blankenbeckler, 2018). The strategic use of 2003). The intention to implement and measure bilingual education as part of an effort to at such a large scale would be difficult if not promote greater inclusion in Mozambique has a accepted and chosen by communities. long history but has always been implemented These initial conversations informed the on a small scale. First developed in 1993, it has design of three situational analyses intended to gradually expanded from a small handful of better understand the program’s sociolinguistic schools to over 500 schools in 2015, still only a context. A language mapping study was small percentage of the total number of over conducted to measure children’s oral language 15,000 primary schools (Benson, 2000, 2004; proficiency (OLP) in each language they Capra, 2013). The bilingual education model is reported knowing. OLP was measured using a transitional, using local languages for semantic fluency measure from a representative instruction in literacy and numeracy in Grades 1- sample of 4,177 Grades 1-3 students in 212 3, with Portuguese as a subject until full schools in Nampula and Zambezia. The transition in Grade 4 (Minsterio da Educação e information gathered assisted policymakers and Desenvolvimento Humano [MINEDH], 2003). education specialists to identify the following: a) The design of Vamos Ler! was initially aimed whether children’s language proficiency to improve reading and writing in the target matched the official language of instruction languages of Emakhuwa and Elomwe in Grades assigned to schools (the study found large 1 and 2 in nearly 3,500 schools in the northern mismatch between self-reports and semantic provinces of Nampula and Zambezia. In early fluency tests, with only about 8% of students joint donor-partner meetings with ministry considered bilingual, and the majority of representatives, it was observed that the students [73%] L1 not matching the school’s program did not fully represent the language of instruction); b) multilingual government’s bilingual transitional approach, as classrooms (approximately 62% of schools were it only focused on the selected L1 languages and linguistically heterogeneous—students with did not include L2 (Portuguese) or L1 numeracy more than one L1); and c) the best choice of instruction in the proposed design. Other language for initial literacy instruction (mapping questions were raised about the language results revealed strong correspondence between 50 Global Education Review 7 (1)

findings and census research, showing correspond to what Rogers (2008) refers to as Echuwabo to have prevalence in southern “‘tipping points,’ where a small additional effort regions, where previously Elomwe may have can have a disproportionately large effect” (p. been used as the selected L1; Nakamura et al., 38). A revised national bilingual education 2018). This study was complemented by policy was released in August 2019 that reflected qualitative research and two additional studies many of the program’s recommendations and that looked at the knowledge, attitudes, and intervention evidence (MINEDH, 2019). practices of community members, teachers, and The Mozambique case illustrates just a few other education stakeholders with regard to of the many sociolinguistic complexities faced in bilingual education in Mozambique (Hua et al., the implementation of L1-based MLE programs, 2017; Shulman, 2017). Findings from the latter including how the program adapted to the assessments showed that parents’ and contextual realities in order to meet program communities’ support for L1-based MLE could goals. Many other psycholinguistic and be vastly improved through more equitable sociopolitical aspects merit investigation that provision of materials for local language have not yet been explored. Weiss (1997) instruction and teacher selection and training to underscores that a program theory (and often improve teacher capacity to teach in the L1 and multiple theories) are “rarely explicit and do not the L2. have to be right” and among other challenges to These studies helped to inform a shift in theory of change design, there may be more than design in many aspects of the program one possible theoretical framework appropriate implementation, including the addition of to the program goal (p. 503). As hypotheses, Echuwabo as a language of instruction in select theories may be drawn from evidence in another schools, and the addition of Portuguese as an context or another set of circumstances, but also area of intervention to improve oral language may be based on assumptions, beliefs, or even development and support transition from the L1 intuition. However, program theory should be to L2 in Grade 4. All four languages (Emakhuwa, developed taking into account the level of Elomwe, Echuwabo, and Portuguese) were complexity, the context, purpose, and evaluation included in an EGRA assessment for evaluation needs particular to each program. If improved of literacy outcomes under the parameters of the learning outcomes is the desired change, there program contract (Turney et al., 2017). Several may be more than one path to achieve this goal, social behavior communications campaigns were or the path may need to take a few alternative or implemented to raise awareness among districts, parallel routes. Learning in the L1 is a core schools, and communities of key L1-based MLE element of academic success. However, this concepts. The program worked with provincial success will be limited, if, among other things, and district authorities to make bilingual policy does not support or is not valued by education mandatory in select schools. The communities, or the L1 is quickly transitioned to program additionally helped fund and promote a the L2, without adequate provision for L2 series of national events and workshops on learning. bilingual education, and in 2019, supported the development of a scope and sequence for the Conclusion: Designing for Complexity remaining 14 national languages in addition to “Bilingual education is a simplistic label for a Emakhuwa, Elomwe, and Echuwabo (Kaplan- complex phenomenon” (Baker, 2011, p 207). Nunes et al., 2019). These latter efforts may Designing for Complexity 51

It is not easy or straightforward to design a France: Association for the Development of strong theory of change. The need for Education in Africa. comparison drives many program theories of Ansah, G. N. (2014). Re-examining the change, which can result in one-size-fits-all fluctuations in language in-education policies in program design. To avoid this trap, flexible postindependence Ghana. Multilingual approaches can better allow for critical changes Education, 4(1), 1-15. at the outset of the program—and along the Archer, A. L., & Hughes, C. A. (2011). way—to reflect new learning as the program Explicit instruction: Effective and efficient progresses. Simple theories often seek to follow teaching. New York, NY: Guilford Press. a linear path of cause and effect, whereas the on- Baker, C. (2011). Foundations of bilingual the-ground realities merit a more complex education and bilingualism (Vol. 79). New York, design where more than one causal strand may NY: Multilingual matters. benefit separate evaluations (e.g., policies may Bartlett, L., Dowd, A. J., & Jonason, C. be required for delivery and implementation of (2015). Problematizing early grade reading: new reading and writing approaches; Rogers, Should the post-2015 agenda treasure what is 2008). These programs might also benefit from measured? International Journal of a design for complexity that depends on Educational Development, 40, 308-314. activating a “virtuous circle” where an initial Bender, P., Dutcher, N., Klaus, D., Shore, J., success creates the conditions for further success & Tesar, C. (2005). In their own language: (Rogers, 2008, p. 38). The flexible approach to Education for all. Education Notes. implementation in Mozambique allowed for Washington, DC: World Bank. critical, small changes that represent one of the Benson, C. J. (2000). The primary bilingual factors that contributed to a “tipping point” in education experiment in Mozambique, 1993 to the policy expansion of bilingual education 1997. International Journal of Bilingual (Rogers, 2008). More guidance and research is Education and Bilingualism, 3(3), 149-166. needed to develop flexible, nonlinear designs for Benson, C. (2004a) Bilingual schooling in L-based multilingual programs that better Mozambique and Bolivia: From experimentation reflect the complexity of the psycholinguistic, to implementation. Language Policy 3, 47-66. sociolinguistic, and sociopolitical contexts, and Benson, C. (2004b). The importance of can allow for more collaborative, sustainable mother tongue-based schooling for educational program implementation. quality. Paper commissioned for the EFA Global Monitoring Report 2005, The Quality Imperative. Paris, France: UNESCO. Retrieved Acknowledgement from I thank Dr. Carol Benson for her valuable http://unesdoc.unesco.org/images/0014/00146 feedback and advice on this manuscript, as well 6/146632e.pdf as for her inspiration and support over the years. Benson, C., & Kosonen, K. (2012). A critical comparison of language-in-education policy and practice in four Southeast Asian countries and References Ethiopia. In T. Skutnabb-Kangas, T., & K. Alidou, H., Boly, A., Brock-Utne, B., & Heugh, (Eds.). Multilingual education and Satina, Y. (2006). Optimizing learning and sustainable diversity work: From periphery to education in Africa – the language factor. Paris, center (pp 111-137). New York, NY: Routledge. 52 Global Education Review 7 (1)

Benson, C. (2013). L1-based multilingual education for all. NABE Journal of Research education and EGRA: Where do they meet? and Practice, 2(1), 1–20. PRAESA Occasional Papers No. 40. Creese, A., Martin, P. W., & Hornberger, N. Benson, C. (2014). Adopting a multilingual H. (Eds.). (2008). Ecology of language. Berlin, habitus: What north and south can learn from Germany: Springer. each other about the essential role of non- Cummins, J. (1979). Linguistic dominant languages in education. In D. Gorter interdependence and the educational et al. (Eds.). Minority languages and development of bilingual children. Review of multilingual education (pp. 11-28). Berlin, Educational Research, 49(2), 222–251. Germany: Springer. Cummins, J. (1984). Bilingualism and Benson, C. & Wong, K. M. (2015). special education: Issues in assessment and Development discourse on language of pedagogy (Vol. 6). Clevedon, UK: Multilingual instruction and literacy: Sound policy and Matters. Ubuntu or lip service? Reconsidering Cummins, J. (2000). Language, power, and Development, 4(1), 1-16. pedagogy: Bilingual children in the crossfire Bialystok, E. (2011). Reshaping the mind: (Vol. 23). Clevedon, UK: Multilingual Matters. the benefits of bilingualism. Canadian Journal Dehaene, S. (2009). Reading in the brain: of Experimental Psychology/Revue Canadienne The new science of how we read. New York, NY: De Psychologie Expérimentale, 65(4), 229. Penguin Group. Blankenbeckler, C. (2017). Vamos Ler! (Let’s Dowd, A. J., & Bartlett, L. (2019). The need Read!) program quarterly progress report 1, for speed: Interrogating the dominance of oral mobilization period: August 12 – December 31, reading fluency in international reading efforts. 2016. Published by Creative Associates Comparative Education Review, 63(2), 189-212. International with USAID support. Retrieved Dubeck, M. M., & Gove, A. (2015). The early from https://dec.usaid.gov/ grade reading assessment (EGRA): Its Canhanga. M. & Banda. M. (2017). theoretical foundation, purpose, and limitations. Education language policy in Mozambique: A International Journal of Educational critical view. International Journal of Development, 40, 315-322 Humanities Social Sciences and Education, Eberhard, D. M., Simons, G. F., & Fennig, C. 4(5): 13-14 D. (Eds.). (2019). Ethnologue: Languages of the Capra International (2013). Evaluation of World. Twenty-second edition. Dallas, Texas: the ministry of education bilingual education SIL International. Retrieved from pilot in Mozambique. Cumberland, ONT, http://www.ethnologue.com Canada: Capra International Inc. Glouberman, S., & Zimmerman, B. (2002). Chen, H. T., & Rossi, P. H. (1989). Issues in Complicated and complex systems: what would the theory-driven perspective. Evaluation and successful reform of Medicare look like?. program planning, 12(4), 299-306. Romanow Papers, 2, 21-53. Clayton, S. (2008). The problem of ‘choice’ Gove, A., & Cvelich, P. (2011). Early and the construction of the demand for English reading: Igniting education for all. A report by in Cambodia. Language Policy, 7(2), 143. the early grade learning community of practice. Collier, V. P., & Thomas, W. P. (2004). The Research Triangle Park, NC: RTI International. astounding effectiveness of dual language Gove, A., & Wetterberg, A. (2011). The Early Grade Reading Assessment: Applications and Designing for Complexity 53

Interventions to Improve Basic Literacy. Hua, H., Noyes, D., Van de Waal, W. & Research Triangle Park, NC: RTI International. Turney, A. (2017). Qualitative study of teacher Graham, B. E., & van Ginkel, A. J. (2014). knowledge, attitudes, & practices (KAP) in Assessing early grade reading: the value and Mozambique, Final Report. Published by limits of “words per minute.” Language, Culture Creative Associates International with USAID and Curriculum, 27(3), 244-259. support. Retrieved from https://dec.usaid.gov/ Hamid, M. O., & Erling, E. J. (2016). Jasińska, K. K., Wolf, S., Jukes, M. C., & English-in-education policy and planning in Dubeck, M. M. (2019). Literacy acquisition in Bangladesh: A critical examination. In R. multilingual educational contexts: Evidence Kirkpatrick, (Ed.). English language education from Coastal Kenya. Developmental science, policy in Asia (pp. 25-48). New York, NY: e12828. Springer. Johnson, D. C. (2013). Language Policy. Hoffman, J. V. (2009). In search of the London: Palgrave Macmillan. “simple view” of reading comprehension. In S.E., Koda, K. (2007). Reading and language Israel & G. G. Duffy (Eds.). Handbook of learning: Crosslinguistic constraints on second research on reading comprehension (pp. 78- language reading development. Language 90). New York: Routledge. Learning, 57, 1-44. Hoover, W. A., & Gough, P. B. (1990). The Koda, K., & Zehler, A. M. (Eds.). (2008). simple view of reading. Reading and writing, Learning to read across languages: Cross- 2(2), 127-160. linguistic relationships in first-and second- Hornberger, N. H. (2002). Multilingual language literacy development. New York, NY: language policies and the continua of biliteracy: Routledge. An ecological approach. Language policy, 1(1), Maddy-Weitzman, B. (2011). The Berber 27-51 identity movement and the challenge to North Hornberger, N. H. (2006a). Frameworks African states. Austin, TX: University of Texas and models in language policy and planning. In Press. T. Ricento (Ed.). An introduction to language Mann, M., & Dalby, D. (2017). A thesaurus policy: Theory and method (p. 24-41). Malden, of African languages: A classified and MA: Blackwell Publishers. annotated inventory of the spoken languages of Hornberger, N. H. (2006b). Voice and Africa with an appendix on their written biliteracy in indigenous language revitalization: representation. New York, NY: Routledge. Contentious educational practices in Quechua, May, S. (2005). Language rights: Moving the Guarani, and Māori contexts. Journal of debate forward. Journal of Sociolinguistics, Language, Identity, and Education, 5(4), 277- 9(3), 319-347. 292. Ministério da Educação e Desenvolvimento Hornberger, N. H. (2009). Multilingual Humano (MINEDH). Instituto Nacional de education policy and practice: Ten certainties Desenvolvimento da Educação (INDE). (2001, (grounded in indigenous experience). Language 2003). A modelo de transição como Teaching, 42(2), 197-211. características de manutenção. Mozambique: Hovens, M. (2002). Bilingual education in MINEDH. West Africa: Does it work? International Minsterio da Educação e Desenvolvimento Journal of Bilingual Education and Humano (MINEDH) (2019). Estratégia Bilingualism, 5(5), 249-266. 54 Global Education Review 7 (1)

nacional de expansão do ensino bilíngue. steps to improve learning outcomes. Prepared Mozambique: MINEDH. for USAID Bureau for Africa. Research Triangle Nakamura, P. R., Carson, C., Davis, E., Rai, Park, North Carolina: Research Triangle N., & Todd, A. (2018). Language Mapping Institute. Retrieved from Study in Mozambique, Final Report. Published https://www.globalreadingnetwork.net/sites/de by Creative Associates International with USAID fault/files/eddata/Language_Use_in_Education support. Retrieved from https://dec.usaid.gov/ _guidance_May_2016_FINAL.pdf Nakamura, P. R., de Hoop, T., & Holla, C. U. Potter, P. & Blankenbeckler, C. (2018). (2018). Language and the learning crisis: USAID Vamos Ler! (Let’s Read!) FY 18 annual Evidence of transfer threshold mechanisms in report: October 2018-September 2018. multilingual reading in South India. The Journal Published by Creative Associates International of Development Studies, 1-19. with USAID support. Retrieved from National Reading Panel. (2000). Teaching https://dec.usaid.gov/ children to read: An evidence-based assessment Ricento, T. (2000). Historical and of the scientific research literature on reading theoretical perspectives in language policy and and its implications for reading instruction. planning. Journal of Sociolinguistics, 4(2), 196- Washington DC: National Institute of Child 213. Health and Human Development. Rogers, P. J. (2008). Using programme Ouane, A., & Glanz, C. (2010). How and theory to evaluate complicated and complex why Africa should invest in African languages aspects of interventions. Evaluation, 14(1), 29- and multilingual education. Hamburg, 48. Germany: UNESCO Institute for Lifelong Ruiz, R. (1984). Orientations in language Learning. planning. NABE Journal, 8(2), 15-34. Piper, B., Schroeder, L., & Trudell, B. Shulman, S (2017). Qualitative study of (2016a). Oral reading fluency and community knowledge, attitudes, & practices comprehension in Kenya: Reading acquisition in (KAP) in Mozambique, Final Report. Published a multilingual environment. Journal of by Creative Associates International with USAID Research in Reading, 39(2), 133-152. support. Retrieved from https://dec.usaid.gov/ Piper, B., Zuilkowski, S. S., & Ong’ele, S. Skutnabb-Kangas, T. (2000). Linguistic (2016b). Implementing mother tongue genocide in education – or worldwide diversity instruction in the real world: Results from a and human rights? Mahwah, N.J.: L. Erlbaum medium-scale randomized controlled trial in Associates. Kenya. Comparative Education Review, 60(4), Skutnabb-Kangas, T., & Heugh, K. (Eds.). 776-807. (2012). Multilingual education and sustainable Perfetti, C. A., & Dunlap, S. (2008). diversity work: From periphery to center. New Learning to read: General principles and writing York, NY: Routledge. system variations. In K. Koda & A. M. Zehler Stanovich, K. E. (1986). Matthew effects in (Eds.). (2008). Learning to read across reading - Some consequences of individual‐ languages: Cross-linguistic relationships in differences in the acquisition of literacy. first-and second-language literacy development Reading Research Quarterly, 21(4), 360–407. (pp. 25-50). New York, NY: Routledge. Tan, L. H., Spinks, J. A., Feng, C. M., Siok, Pflepsen, A. (2015). Planning for language W. T., Perfetti, C. A., Xiong, J., Fox, Peter T., & use in education: Best practices and practical Gao, J. H. (2003). Neural systems of second Designing for Complexity 55

language reading are shaped by native language. UNESCO. Retrieved from Human Brain Mapping, 18(3), 158-166. https://undocs.org/E/2019/68 Tinsley, T. (2013). Languages: the state of United Nations (2008). United Nations the nation. Demand and supply of language declaration on the rights of indigenous peoples. skills in the UK. London, UK: British Academy. New York: United Nations. Retrieved from Tollefson, J. W. (Ed.). (2002). Language https://www.un.org/en/genocideprevention/do policies in education: Critical issues. New York, cuments/atrocity- NY: Routledge. crimes/Doc.18_declaration%20rights%20indige Trudell, B., & Piper, B. (2013). Whatever the nous%20peoples.pdf law says: Language policy implementation and USAID (2012, April). USAID 2011 education early-grade literacy achievement in Kenya. strategy: Technical notes. Washington, DC: Current Issues in Language Planning, 15(1), 4- USAID. Retrieved from 21. http://pdf.usaid.gov/pdf_docs/PDACT681.pdf Turney, A., Noyes, D., Dhliwayo, T. M., USAID (2018, November). USAID education Machkasov, Y., Van de Waal, W. & Hua, H. policy: Program cycle implementation and (2018). Effectiveness Evaluation Baseline operational guidance. Washington, D.C.: Report – Mozambique. Retrieved from USAID. Retrieved from https://dec.usaid.gov https://www.usaid.gov/sites/default/files/docu UNESCO (1953). The use of the vernacular ments/1865/USAID_Education_Policy_Progra languages in education. Monographs on m_Cycle_Implementation_and_Operational_G Foundations of Education, No. 8. Paris, France: uidance_FINAL.pdf UNESCO. United States Government [website]. (n.d.). UNESCO (2003). Education position paper: Resources retrieved in October, 2019, from Education in a multilingual world. Paris: https://www.grants.gov/learn-grants/grant- UNESCO. Retrieved from making-agencies/agency-for-international- https://unesdoc.unesco.org/ark:/48223/pf0000 development.html and https://beta.sam.gov/ 129728 Walter, S., & Benson, C. (2012). Language UNESCO (2014). Teaching and learning: policy and medium of instruction in formal achieving quality for all (Education for All education. In B. Spolsky (Ed.), The Cambridge Global Monitoring Report 2013/4). Paris, handbook of language policy (pp. 278-300). France: UNESCO. Retrieved from Cambridge, UK: Cambridge University Press. http://unesdoc.unesco.org/images/0022/00225 Wardhaugh, R. (2011). An introduction to 6/225660e.pdf sociolinguistics (Vol. 28). Hoboken, NJ: John UNESCO (2016). Education for people and Wiley & Sons. planet: Creating sustainable futures for all Weiss, C. H. (1997). How can theory-based (Education for All Global Monitoring Report evaluation make greater headway? Evaluation 2013/4). Paris, France: UNESCO. Retrieved Review, 21(4), 501-524. from Weiss, C. (1998). Evaluation: Methods for https://unesdoc.unesco.org/ark:/48223/pf0000 studying programs and policies. Englewood 245752 Cliffs, NJ: Prentice Hall. UNESCO (2019). Report of the secretary- Wiley, T. G. (2012). A brief history and general. Special edition: Progress towards the assessment of language rights in the United sustainable development goals. Paris, France: 56 Global Education Review 7 (1)

States. In Language Policies in Education (pp. 73-102). NY: Routledge. Wolf, M. (2008). Proust and the squid: The story and science of the reading brain. Cambridge, UK: Icon. Ziegler, J. C., & Goswami, U. (2006). Becoming literate in different languages: Similar problems, different solutions. Developmental Science, 9(5), 429-436.

About the Author Corrie Blankenbeckler is a senior technical advisor in the Education Division at Creative Associates International. She has over fifteen years' experience in developing and managing formal and non-formal education programs in diverse environments in the Middle East/North Africa, Sub-Saharan Africa, and Asia. In her current role, she leads in program design and provides technical quality assurance and support for Creative projects around the world. Since 2016 she has provided project direction and technical support for the bilingual early grade reading project in Mozambique. She has an M.A. in international education development from Teachers College of Columbia University, and has deep expertise in mother-tongue based multilingual curriculum design.