Multilingualism and the International Baccalaureate Diploma Programme (IBDP): An Analysis of Language Learning in the IBDP in Light of the ‘Multilingual Turn’

This article analyses various International Baccalaureate policy documents to establish whether the International Baccalaureate Diploma Programme has undergone what is described as a multilingual turn. After defining multilingualism and the multilingual turn, it outlines three main implications of what might be considered this paradigm shift for educational policies related to curriculum, pedagogy, and assessment. These implications are used as a framework for the policy analysis to follow, which argues that International Baccalaureate Diploma Programme approaches to language learning are underpinned by a monolingual bias that might delegitimate the knowledge and experience of multilingual learners.


Introduction
The International Baccalaureate (IB) Diploma Programme (DP) 'was created in 1968 to provide a truly international education for students preparing for university' (Abrioux and Rutherford, 2013: 79) and has expanded (as of September 2019) to 3,421 schools offering the Diploma in 157 countries around the world (IB, 2020a). Throughout its history, language learning has been an important part of the Diploma, with students required to study at least two language subjects, at either Higher Level (HL) or Standard Level (SL), out of a total of six main subjects, in groups described as Groups 1-6 (in addition to the three core components: Extended Essay, Theory of Knowledge and Creativity, Activity, Service (CAS) (IB, 2021a). IBDP students are linguistically diverse with 192 'first languages' reported amongst May 2020 examination candidates (IB, 2020: 2). These students include those learning through an additional language in their home countries as well as more

The Multilingual Turn
The publication in 2014 of two books entitled The Multilingual Turn (May, 2014;Conteh and Meier, 2014), despite being 'conceived independently on opposite sides of the globe', adds strength, as Conteh and Meier (2014: 3) put it, 'to the argument that the multilingual turn may indeed be here'. But what is the multilingual turn? 'Multilingual' here does not refer to those using three or more languages as it is sometimes used (with 'bilingual' referring to users of two languages) (see, eg, Aronin and Hufeisen, 2009: 4;Nguyen, 2014: 9); rather, the multilingual turn represents, as Meier (2016) argues, a fundamental shift in the way we conceptualise language, learners and language learning. In this way, the multilingual turn represents a paradigm shift and, like all paradigm shifts, 'involve[s] the adoption of a new outlook or paradigm' (Richards and Schmidt, 2010: 418). This new outlook has brought 'modifications of rules and assumptions that underlie foreign language teaching' (Lafford, 2000: 723).
Such modifications arise from the insight that languages do not exist within the bilingual mind 'as two solitudes that should be kept rigidly separate' (Cummins, 2008: 65), as developments in various fields have shown. Neuroscientists, for example, have found that, rather than existing separately, 'both of a bilingual's languages are active most of time' (Grant et al, 2019: 48) with 'information in all of the languages they know . . . momentarily active' when multilinguals process words from one language (Kroll et al, 2013: 102). Sociolinguists have argued that 'there can be little doubt that the notion of language should be viewed as a construction rather than an objectively observable coherent entity' (Møller, 2019). Educationalists meanwhile have shown that, despite the pervasive monolingual bias in most education systems (see, for example, Auer, 2007;Otwinowska, 2017;Creagh, 2017;Escobar, 2016), students' use of multiple named languages in the same discourse as 'codes integrated into a hybrid language' (Michael-Luna and Canagarajah, 2007: 58) is so common that it is 'something that is natural for bilinguals to do' (Turnball and Dailey-O'Cain, 2009: 1).
As scholars have sought to understand the implications of such developments and the significance of this paradigm shift for education, there has been a 'terminological proliferation' (May, 2014: 1), with the emergence of a range of terms -'polylingual' (Jørgensen, 2008), 'metrolingual' (Pennycook, 2012: 18), 'codemeshing' (Canagarajah, 2011), 'translanguaging' (García, 2009), 'heteroglossia' (Blackledge et al, 2014) -all sharing, albeit with subtly different emphases, the understanding that languages do not exist as separate essentialist systems; rather, each individual draws upon a single linguistic repertoire consisting of what the monolingual bias names as "languages". This represents, to borrow Bernstein's (1971) terminology, a weakening of the classification between languages. To acknowledge and grapple with the implications of this weakened classification is to undergo a multilingual turn.
However, most schools have not yet made this turn, and the 'monolingual habitus' (Gogolin, 1997; still holds strong (see also García and Woodley, 2016;Hajek and Slaughter, 2015). This situation demands that we ask, as García (2007: xiii) does, 'What would language education look like if we no longer posited the existence of separate languages?' Before examining IBDP documentation, we will first briefly consider this question and the implications of the multilingual turn for language education policies.

Implications of the multilingual turn for language education policies
The implications of the multilingual turn for language education are here explored in the three areas that are Bernstein's (1971: 47) three message systems of schooling: curriculum, pedagogy, and evaluation. Although these areas have, as Rizvi and Lingard (2010: 93-94) point out, been neglected in the field of education policy studies, 'it is these three message systems which frame the core of teachers' work, which inform their logics of practice.' Bernstein 'was also interested in how these message systems linked to the broader society and culture' (Lingard et al, 2016: 9), using the term 'educational knowledge code' to refer to 'the underlying principles which shape' these three message systems (1971: 47). So, to imagine educational policies underpinned by principles of multilingualism, we can consider the implications of the paradigm shift for curriculum, pedagogy, and evaluation, and then later analyse IB policies that guide these message systems, considering the educational knowledge code they seem to draw upon.
To broadly summarise the implications in these three areas, we would expect educational policies that have undergone a multilingual turn to: 1. envisage a curriculum that does not rely on simplistic concepts like 'mother tongue', 'second language' or 'native speaker', and simplistic assumptions about the relationship between language and culture; 2. promote pedagogies that open up translanguaging spaces; 3. ensure fair assessment (evaluation) of multilingual students.
Each of these implications is explained below, and in the following section will be used to frame the subsequent IB policy analysis.

1.
Policies will envisage a curriculum that does not rely on simplistic concepts like 'mother tongue', 'second language' or 'native speaker', and simplistic assumptions about the relationship between language and culture.
The multilingual turn complicates terms and concepts previously taken for granted, and we would expect this to be reflected in policies. The first concept that becomes inadequate in a multilingual reality is that of the 'native speaker' which, as Davis (2003: 2,7) has proposed, is both ambiguous and exclusionary. Canagarajah (1999: 77) has argued that, in addition to excluding many types of language user such as childhood bilinguals and members of postcolonial communities, this simplistic term is often used to portray one type of language use as superior or normative, which has 'non-linguistic roots [and] non-pedagogical results'. Related to this is the term 'mother tongue', which is so difficult to pin down that even censuses define it in different ways (Makoni & Pennycook, 2007: 12). Despite this, 'the gendered and affectively charged kinship concept of the unique "mother tongue" . . . established the idea that having one language was the natural norm' (Yildiz, 2012: 6), which is not borne out by the multilingual reality. Indeed, as Piccardo and Aden (2014: 238) argue, 'thinking in terms of mother tongue, first, second and foreign language . . . automatically implies separation and hierarchy', denying the blurred boundaries of multilingualism. Instead of strongly classified dichotomies such as 'native speaker' or 'mother tongue', and 'non-native speaker' or 'second language' that privilege monolingualism (Dewaele, 2018), the multilingual reality demands that policies reject such inadequate concepts. Yildiz (2012: 2) also questions another assumption implied by the concept of 'mother tongue': According to this paradigm, individuals and social formations are imagined to possess one "true" language only, their "mother tongue", and through this possession to be organically linked to an exclusive, clearly demarcated ethnicity, culture, and nation.
This assumption, that a language is intrinsically connected to a particular culture, is challenged by the multilingual turn. English is perhaps the most obvious illustration of this: given that a large majority of English speakers in the world are multilingual users who live in places where English is not an official language, it no longer makes sense to speak of 'English speaking cultures' (Baker, 2009). In this, English is perhaps unique, but in the cases of other named languages it is certainly also true that 'there is a polyphony of cultures within which the target language is spoken as well as polyphony within each of those cultures' (Crawford and McLaren, 2003: 136). As well as this diversity, García (2009: 79) notes that multilingualism allows individuals to create hybrid cultures that combine elements of different cultural systems, again questioning the assumed simplistic link between language and culture.

2.
Policies will promote pedagogies that open up translanguaging spaces.
One clear implication of the multilingual turn for educators is that pedagogies teachers use should reflect the multilingual reality and use this as a resource for learning. We would expect such policies to support translanguaging pedagogies that 'intentionally enact and invite multilingual contributions' (Hopewell, 2017: 165). Translanguaging is defined by Otheguy, García and Reid (2015: 281) as 'the deployment of a speaker's full linguistic repertoire without regard for watchful adherence to the socially and politically defined boundaries of named (and usually national and state) languages'. While a policy might not use terms like 'translanguaging', it should encourage the use of pedagogies that allow students' use of their full linguistic repertoires, thus treating languages as social resources and respecting students' agency as to which parts of their linguistic repertoire they use in their learning (Cenoz and Gorter, 2014) Furthermore, García (2005: 226), building on work by Cummins (2000), notes that an effective pedagogy for multilingual learners is based on two principles of social practice (which includes dimensions such as interaction, involvement, and collaboration) and social justice (which includes equity, rigor, and language tolerance). To build on this, I borrow Li's (2011) concept of a translanguaging space, a space in which multilinguals are able to draw on their multiple linguistic resources and by doing so open up spaces for new transformations, learnings, identities, and so on. It seems that certain pedagogical approaches -those that are underpinned by principles of social practice and social justice -would be more likely to open up translanguaging spaces. Constructivist teaching approaches that encourage collaboration and independent inquiry are more likely to enable students to draw on their full linguistic repertoire, for instance. On the other hand, other pedagogical approaches would be more likely to close down translanguaging spaces. As Canagarajah (2013: 185) warns, 'traditional classroom relations and pedagogical approaches have the potential to stifle the dispositions and competences which students bring with them'. A classroom with a strict English Only rule, for example, would close down translanguaging spaces. When it comes to analyzing IB policies, it is important to consider whether the pedagogies they promote are rooted in social justice and social practice, and open up translanguaging spaces.

3.
Policies will ensure fair assessment of multilingual students.
In the field of multilingual education, the development of assessment practices that fairly and accurately test multilingual students is often seen as an urgent and important challenge (García, 2009: 378;García and Wei, 2013: 119) given that most standardized testing is predicated on the assumption of monolingualism (Shohamy, 2011: 418;Lopez et al, 2017: 1). These assessment practices disadvantage multilingual students who, unlike their monolingual peers, cannot use their full language repertoires to demonstrate understanding (García and Kleyn, 2016: 24) and will be penalized for reasons of language proficiency rather than content knowledge, undermining the tests' reliability and validity (Menken, 2008: 147).
To move away from this monolingual bias and integrate the multilingual reality, it is argued that assessment should be redesigned around 'more holistic approaches that consider language-asresource and promote the use of whole linguistic repertoire' (Dorter and Cenoz, 2017: 231). The best way of achieving this is not yet clear, but various possibilities have been proposed. Shohamy (2011) places such approaches on a continuum, at one end of which are assessment practices that provide questions or prompt information in multiple languages but require a response in one language. The bilingual tap approach described by García (2009: 263) would be placed here on the continuum. At the other end are approaches that permit students to mix named languages and score such hybrid language use without penalty. An example here would be the mathematics examination trial described by Lopez et al (2017), in which students can respond in Spanish, English, or a mixture of both.
Although, as these examples show, it is possible for high-stakes examination-style assessment to undergo a multilingual turn, some have argued that alternative assessment methods might better suit multilingual learners. García (2009: 267), for instance, claims that 'the best way to assess bilingual students is for teachers to observe and listen to their students, and to record these observations systematically over long periods of time'. Various assessment tools have been developed that could support such observations, such as the Multilingual Assessment Instrument for Narratives (Gagarina et al, 2012), the Pluricultural and Plurilingual Scales (Little et al, 2018), and the European Language Portfolio (CoE, 2020).
To summarise, this section has identified various implications of the multilingual turn for educational policies related to curriculum, pedagogy, and assessment, which will be used to frame the analysis of IB policies that follows.

IBDP Policy Analysis
To guide the selection of documents to be analysed in this study, Hall and Hord's (1987: 183) definition of policy is taken as a starting point: 'an explicit statement . . . published in the records of the organization [and] officially sanctioned by authority' that 'reflects or directs the procedures, decisions, and actions of an organization and the individuals within it'. The IB has not published a single policy about language learning but rather a collection of interrelated documents -which Ball (2006: 48) might refer to as a 'policy ensemble' -that direct 'the procedures, decisions and actions' of schools relating to languages. These documents will now be analysed to consider whether they reflect the implications (for curriculum, pedagogy, and assessment) of the multilingual turn outlined above.

1.
Policies will envisage a curriculum that does not rely on simplistic concepts like 'mother tongue', 'second language' or 'native speaker', and simplistic assumptions about the relationship between language and culture.
As explored above, concepts such as 'first language', 'mother tongue' and 'second language' become inadequate in our multilingual reality. This seems to be something that the IB has been aware of since its earliest days, as this 1967 IB document (in Tosi, 1991:90) suggests: The most complicated problem arises from the different linguistic background of the students. By choice or by necessity, many of them use a language of instruction which is not their mother tongue; at the same time their mother tongue may be taught to them as a foreign language or they may use it only at home or in private lessons. Consequently, instead of talking of 'mother tongue' on the one hand and 'foreign languages' on the other, it seems more desirable to call the first means of expression (of which a pupil should have complete command at the end of his secondary school studies) Language A, and the second (of which he should have an adequate knowledge) Language B.
Here, the IB recognizes the problem that others (eg Pennington, 2012 ;Carder, 2006) later identified: the complex linguistic background of many IBDP students means they do not fit into neat dichotomous categories such as 'mother tongue' and 'foreign language'. However, it is unclear whether the creation of these new dichotomous labels of 'Language A' and 'Language B' has solved this problem, especially as IB policies continued to use such terms that it had thought to be inadequate, as these examples illustrate (emphasis my own): The Language A1 programme is a literature course studied in the "first language" of the student or the language in which the student is most competent (IB, 1999: 3) Students should be given the opportunity to study their best language/mother tongue (IB, 2009a: 22).
[A school's language policy must] outline how students are to learn at least one language in addition to their mother tongue (IB, 2015d: 35).
This use of terms in policies that the IB itself has acknowledged to be inadequate in light of students' multilingualism is also evident in one document entitled Language and Learning in IB Programmes, where IB (2011) affirms 'multilingualism as a fact' (p10), mentions 'the complex conceptualization of how individual language profiles are constructed' (p5), and even refers to Orwell in claiming notions of 'native speaker' and 'foreign language' are ideological 'grand generalisations' (p5). However, this is undermined by repeated references throughout the same document to 'mother tongue', which it defines as 'the language learned first and/or the language identified with as a "native speaker". . . [and] the language that the student uses at home and/or outside the classroom environment' (p16). The term 'mother tongue' is not rendered any less troublesome by this definition -a 'grand generalisation', we might say -that seems to ignore students who, for example, might use various named languages in different contexts 'outside the classroom' or who learn multiple named languages alongside one another. It is, however, notable that the use of such terms has become less frequent in more recent policies. For instance, whereas an older policy like IB (2009b: 33) titled Group 2 (also known as Language B) as 'Second Language', the current equivalent policy (IB, 2018c: 262) terms it 'Language Acquisition'. Similarly, early policies described Group 1 (also known as Language A) as study of the 'first language', 'mother tongue', 'native language', 'home language' or 'the language of the environment to which the student has been exposed from an early age' (IB, 1999: 3), but this later changed to 'best language' and the language in which the student is 'highly competent . . . whether or not it is their mother tongue' (IB, 2010). The most recent equivalent document does not in fact distinguish between the study of languages in Groups 1 and 2, stating simply that 'students study two modern languages' (IB, 2019c). The reduction in its use of terms such as 'mother tongue' and 'foreign language' and, perhaps, in not differentiating the two, suggests a weakening of the classification between Groups 1 and 2 languages. So, is the IB, as the multilingual turn, blurring the boundaries between these language groups?
Digging deeper, the strength of classification between Languages A and B can be identified by analyzing how IB Subject Guides (guidance for teachers relating to curriculum content and assessment) describe the nature of each subject. Early policies were explicit that Group 2 subjects constituted language acquisition courses 'from the comparatively elementary, practical usage at ab initio level to the sophisticated usage of the near-native (or bilingual) speaker' (IB, 2002a: 3), whereas the Group 1 Language A subject was intended for students studying their 'first language'. In fact, a Subject Brief stated explicitly that Language A subjects 'are not language-acquisition courses' (IB, 2010), clearly seeking to insulate the subjects from each other, strengthening the classification. However, as with the changing terminology used, the way in which policies describe the nature of the subjects has become vaguer and arguably less strongly classified. The current Language A Subject Guide, for instance, states that Group 1 language courses are 'designed for students from a wide variety of linguistic and cultural backgrounds' who 'will often have varied language profiles and may be multilingual' (IB, 2019d: 6, 11). The Language B Subject Guide describes the subject as a 'language acquisition course designed for students with some previous experience of the target language' (IB, 2018d: 6). The clarity of previous Subject Guides -which even included multiple pages outlining which language courses are suitable for which kinds of student (IB, 2002b: 3-6) -has been replaced by more general descriptions that seem to blur the distinctions between the two language Groups. If, as Bernstein (1971: 49) put it, classification 'refers to the degree of boundary maintenance between contents', IB policies seems to be less concerned about maintaining the boundaries between Languages A and B, at least in terms of terminology and subject descriptions, as it was in the past.
If this is the case, such a weakening of classification would be unusual in specialised collection codes like the IBDP curriculum (for Bernsteinian analysis of the IBDP curriculum see Cambridge, 2010Cambridge, , 2011Cambridge, , 2012 which, Bernstein (1971: 55) argues, 'abhor mixed categories and blurred identities, for they represent a potential openness, an ambiguity . . .' It would also be unusual, given that most education models 'have constructed the two subjects of L1 [first language] and FL [foreign language] as discrete areas of knowledge' (Pomphrey and Burley, 2009: 424). In fact, by applying Evans' ideas below, we see that the fundamental identity of the IB language courses remains more strongly classified than the changing descriptions of the subjects suggest.
For Evans (1993: 51), the notions of boundary and identity are 'complementary' and 'at the heart of the discipline'. Strong classification creates 'powerful subject identities' (Evans, 1988: 163) as it is from 'firm, clear boundaries' that disciplines 'derive their identities' (Evans, 1993: 51-2). But Evans notes that knowledge of these boundaries is secondary to knowledge of a discipline's core. It is by 'think [ing] in terms of a core, particularly a materials core' that a group works towards 'achieving group identity' (Evans, 1993: 170): The key to knowing your discipline (your group) in this perspective is knowing your core, your essence, your principle of coherence. You also need to know your boundaries, but this is secondary because, if someone else's core is different, they are by definition beyond your boundaries. (Evans, 1993: 161). Evans' interest in the core of disciplines related mainly to the teaching and learning of English at universities, where he noted there had been a deep historical split 'between the notion that the core material of [the subject English] is language and the notion that it is literature ' (1993: 170). He sees this distinction as arbitrary ('it should be possible to switch easily between one and the other') but as one that generates the discipline's identity. And so it is for the IBDP language courses: despite the use of vaguer terms and descriptions that seem to blur the boundaries between Languages A and B, IB policies are clear in establishing distinct cores for each subject. In the Language A course, 'students will focus exclusively on literary texts' and 'explore the nature of literature, the aesthetic function of literary language and literary textuality, and the relationship between literature and the world' (IB, 2019e). By contrast, the Language B course emphasizes communication through receptive and productive skills, with no required literature study at Standard Level, and two literary works required at Higher Level (IB, 2018d). However, the policy is clear that literature is studied in Language B HL as a vehicle for developing students' receptive and productive skills, and states that 'it must be emphasized that literary criticism is not an objective of the Language B course; literary criticism lies within the remit of the DP studies in language and literature courses' (IB, 2018d: 22). So, the IBDP policies are explicit in establishing that the core of Language A is literature and the core of Language B is communication. The creation of clear, distinct cores strengthens the classification between Languages A and B, and it is further strengthened by rules that state, for example, that literary works by the same author studied in a student's Language B HL course cannot also be studied in Language A (IB, 2019e: 22).
So, we have seen that although IB appears to be avoiding use of some troublesome terms and dichotomies in recent policies, the way in which languages are treated in the IBDP curriculum seems more strongly classified and draws upon dichotomous concepts that, as explored above, have become inappropriate following the multilingual turn.
IB policies also sometimes present culture in a straightforward manner, unquestioningly and inextricably linked to language, in a way that is unsuitable given the implications of the multilingual turn outlined above. For example (my emphasis): [T]he only way to appreciate another language or culture is to be confident first with your own. Language learning, and learning about different cultures through language, plays a pivotal role in the programme. (IB, 2015d: 6) An IB education creates teaching and learning communities and opportunities that help students increase their understanding of language and culture (IB, 2015d: 11) The Language B Subject Guide similarly promotes what Baker (2009: 256) has termed 'the traditionally conceived target language-target culture relationship', suggesting that particular cultures somehow belong to particular languages, as these quotations illustrate: [L]anguage B students must study authentic texts that explore the culture(s) of the target language (IB, 2018d: 17).
Reading literature in the target language can be an enjoyable journey into cultures where the target language is spoken (IB, 2018d: 22).
. . . provide the student with opportunities to demonstrate his or her understanding and appreciation of the target language culture(s) (IB, 2018d: 45).
As teachers make practical curriculum decisions, they will likely be left with questions regarding the adequacy of the term 'target language culture(s)'. Could a Spanish B teacher include a text written by a Puerto Rican writer in East Harlem? Would the novels of Amy Tan, born and raised in California, written in English and exploring the experience of Chinese-Americans, be considered to be 'texts that explore the culture(s) of the target language' in an English B class? What about Whale Rider, a New Zealand novel written in English that explores, in part, a Maori community's efforts to preserve their indigenous language and culture against the encroachment of the English language?
This oversimplification of the relationship between language and culture misrepresents complex multilingual realities. More than this, it might actually undermine the IB's aim of promoting international mindedness. Kumaravadivelu (2008: 173) argues that an exclusive focus on supposed target language culture(s) in language education might lead students to falsely believe that knowledge of the culture presented in their language classes is enough 'to prepare them to face the challenges posed by economic and cultural globalization'. This is especially the case when a curriculum encourages 'passive reception' of cultural knowledge (Kumaravadivelu, 2008: 181), which might result from the IB Language B course's 'enjoyable journey[s]' of 'understanding and appreciation of the target language culture(s)'. Instead, Kumaravadivelu (2008: 174) calls for language curricula that demand students critically reflect on 'clusters of cultures' that represent complex realities and are relevant to students' interests and contexts.
From this analysis, it seems clear the curriculum envisaged in IBDP policies sometimes draws upon simplistic concepts, even if the terms used to refer to these have become less problematic over time, and, as such, it cannot be said to have undergone a multilingual turn.

Policies will promote pedagogies that open up translanguaging spaces.
A variety of policies set out the IBDP's vision of its distinctive pedagogy, underpinned by six principles (IB, 2017b: 6) which state that teaching will be: ○ based on inquiry ○ focused on conceptual understanding ○ developed in local and global contexts ○ focused on effective teamwork and collaboration ○ differentiated to meet the needs of all learners ○ informed by assessment (formative and summative) As noted above, an analysis of the pedagogical principles envisaged in the IBDP policies should consider whether these approaches are more likely to open or close translanguaging spaces. A number of these key principles have the potential to open up such spaces. Inquiry based learning, for example, could leverage student agency and allow students to independently select resources and learning materials to use their whole linguistic repertoire, as Garza and Arreguín-Anderson (2018) found in their study of translanguaging in an inquiry-based science class. Similarly, collaboration is one key dimension of the social practice principle that García (2005: 226) argued should underpin effective pedagogy for multilingual learners, and studies such as that of Mbirimi-Hungwe (2016) have shown that allowing students to use their whole language repertoires in group discussions helps build understanding. Furthermore, the statement that pedagogies used in the IBDP should contextualise learning so that it 'places an emphasis on students processing new information by connecting it to their own experience and to the world around them' (IB, 2015b), as well as 'activate prior learning, including that learned in other languages' (IB, 2015a), is clear encouragement for teachers to open up translanguaging spaces.
On the other hand, it must be acknowledged that such opportunities could be missed if, for example, teachers lack knowledge about how to activate students' prior learning in other languages, as studies have found happens with teachers in other contexts (De Angelis, 2011;Haukus, 2016). Similarly, it would be possible for group work or inquiry-based learning to be organized in ways that close rather than open translanguaging spaces if teachers, for example, insist on English Only approaches.
Some more detailed pedagogical guidance is provided in other polices. IB (2011: 1) emphasizes the important of pedagogies taking account of students' 'multilingual language profile[s]' and notes that students might progress differently in languages that they use for different purposes, reflecting the concept of a linguistic repertoire used as a social resource and corresponding with Cook's (2016) theory of multicompetence. The policy describes a 'common pedagogy for language and learning' (p28) that draws on the work of Cummins and highlights the importance of building on students' background knowledge (including knowledge 'encoded in their mother tongue or other languages'(p29)) and the essential role of 'social and emotional conditions for learning that value all languages and cultures' (p30), part of the social justice principle outlined above. In these ways, IB (2011) adds more detail to principles that are mentioned briefly elsewhere and encapsulates a pedagogical vision that could open up translanguaging spaces. It is unclear, however, what influence this policy has when, as Carder (2013: 93) notes, 'there is no requirement for schools to follow' it and 'there are limited sanctions on schools for not adhering to good practice'. Furthermore, as this policy is not directly linked to specific curricula or higher stakes assessment guidance, its influence over pedagogy might be less powerful than other documentation such as Subject Guides, which do not contain this clearly described pedagogical vision.
An aspect of IB pedagogy that is probably more familiar to teachers is the Approaches to Learning (ATLs) referenced in many policies, including Subject Guides. The IB emphasizes five ATLs -thinking skills, communication skills, social skills, self-management skills, research skills -and policies state that teachers should teach 'with the development of ATL skills in mind [which] impacts significantly on the wider pedagogical approach adopted' (IB, 2015c). It is evident that a pedagogy suited to multilingualism -one that, as outlined above, enables students to use their agency in drawing from their whole linguistic repertoire and is underpinned by social justice and social practice -would require the development of these ATLs, particularly social skills, communication skills, and self-management skills. While it is possible that a pedagogy that emphasizes development of these ATLs would open up translanguaging spaces, and it is hard to see how such a space could be opened without them, it is, again, not necessarily the case that this will happen. A teacher would have to help students to direct these skills towards their whole linguistic repertoire by, for instance, normalizing multilingual communication skills (Weber and Horner, 2012: 201) or building students' self-management skills so they can make decisions about which parts of their linguistic repertoire to draw on when learning (Velasco and García, 2014).
So, it seems that many aspects of the pedagogical vision outlined in IBDP policies have the potential to open up translanguaging spaces and reflect the multilingual reality; the extent to which such possibilities are realized in the everyday practice of teachers is, however, unclear.
3. Policies will ensure fair assessment of multilingual students.
Most assessment in the IBDP relies on external, large-scale standardized examinations, but these are combined with elements of internal assessment in many subjects, including speaking tests and written assignments, graded by teachers and externally moderated by the IB. Brown (2002: 65) believes that the flexibility of internal assessments means they are more able to reflect cultural differences of international students, but the same cannot be said of linguistic differences. In contrast to the more innovative multilingual assessment approaches outlined above, mixing of named languages is not permitted in internal or external assessment. In this way, as Tosi (1991: 89) observes, the IB 'adopts monolingual not bilingual education'. IBDP policies state that assessments can be taken in the IB's 'working languages' of English, French, or Spanish, with limited subjects available in German, and Theory of Knowledge assessments also available in Chinese (IB, 2018c: 72, 361). Whilst it is true that the availability of these response languages provides multilingual students with more choice than other international assessments offered by competitor organisations, this approach is still underpinned by monolingual assumptions about the existence of languages as separate systems.
Rather than other multilingual assessment approaches, the IBDP's approach to external assessment relies on the translation of assessment instruments into a small number of languages. García (2009: 265;2010: 206) outlines a number of concerns with this approach, questioning the validity of translated assessments due to the difficulty of maintaining test equivalence. A recent example illustrates such challenges: in September 2018, IB issued a clarification amending the translation of two terms, and further explaining the intended meaning of one term, used in the German translation of Theory of Knowledge essay prompts (IB, 2018b). Further difficulties arise when there is a mismatch between the languages of instruction and assessment response languages: 'translations are only appropriate if the student has been taught through the language of the test' (García, 2009: 265). A student who was taught in English might feel more confident responding to examination questions in Spanish, for example, but would likely lack knowledge of specialist subject terminology in Spanish, a problem that could be solved if the student were encouraged to use English terminology alongside Spanish explanations. Ballantyne and Rivera's (2014: 118) study of 300 IB schools found nine schools in which examinations were not conducted in the school's language of instruction (for example, a school teaching through Turkish and examining through English), but it is unclear how common a mismatch of teaching and examination languages is among individual students.
The exceptions to IB's approach of translation of examination instruments into a small number of named languages are in the Language A and Language B assessments, which, whilst they assess the same set of outcomes, are developed from the beginning in the target languages, a process García (2009: 365) terms transadaption. Another difference is that students can be examined in a wider number of named languages in the Language A course. IB policies outline that schools can request on behalf of students any language to be studied and examined; students in 2019 were examined in 55 languages, including 20 students in Sesotho and 14 students in Burmese, for example (IB, 2019a). IB's commitment to offering a wide range of named languages in Group 1 is certainly laudable and demonstrates the organisation's interest in multilingualism; it is, however, problematized by the complicated issues surrounding how a 'language' is defined.
To be studied and taught as a Group 1 subject, a language must have a written literature, which in itself eliminates certain languages or dialects, and disadvantages some students. This has been an issue for some students since early in the IBDP's history: Peterson (1987: 65), for instance, describes denying a West African student's request to be examined in Wollof because of the lack of a written literature at that time. The later addition of a requirement to study texts translated into the language of study, as outlined in policies concerning Language A courses (eg IB, 2019d), has further limited the number of possible languages, disadvantaging students from developing countries in particular (Tosi, 1991: 90). These restrictions notwithstanding, IB policies frame the ability for a student to request any language as fulfilment of a right to multilingualism (eg IB, 2011: 19), but this entitlement apparently does not extend to Group 2 subjects, where the availability of examinable languages is determined by 'market demands' (IB, 2019g) and requests for additional examinable languages require justification from a group of schools showing demand for the language from 'a minimum of 5 schools and a potential 75-100 candidates per session for a forecast period of 3-5 years' (IB, 2019f). In these ways, assessment procedures and regulations outlined in IB policies privilege only narrow understandings of multilingualism (ie a student with very strong proficiency in a language with a well-developed literature that includes translated texts) above others.
The final aspect of IBDP policies regarding assessment worth noting is the regulation that a student is permitted to use a 'bilingual/translation dictionary' for examinations in Groups 3 -6 subjects 'if the response language of the examination is not the best language of the candidate' (IB, 2018c: 401). Putting aside issues with the term 'best language', research into the effectiveness of assessment accommodations such as these for multilingual students is inconclusive (Abedi et al, 2004). García (2009: 265) observes that most education systems rely on accommodations such as test translation and dictionary use 'because of little interest in developing appropriate bilingual assessment'. In the future, if it engages more with the multilingual assessment possibilities outlined above, the IB might move towards developing more appropriate assessment approaches that do not demonstrate a monolingual bias.

Conclusion
As the preceding discussion has shown, IBDP policies promote a particular vision of language learning that seems to be based on a narrow understanding of multilingualism. As Tate (2016: 30) and Tosi (1991: 93) observe, this model is well-suited to a monolingual student, who can study a little of an additional language at school and then think of him/herself as multilingual. This is what García and Otheguy (2020: 18), drawing on Fishman (1977), refer to as a kind of 'elite bilingualism': When elite monolinguals develop as bilinguals, they most often do so in school, where they are taught what is labeled as a second language, to be used completely separately from what is called their first language or mother tongue.
If the IBDP curriculum model effectively enables elite monolinguals to become elite bilinguals/ multilinguals, the curriculum, pedagogy, and assessment practices it promotes might not be suitable for developing the language learning or reflecting the experiences of those who are already multilingual (Tosi, 1991: 93). In this way, the policies seem to understand multilingualism through what has variously been termed 'a monoglossic lens' (García and Woodley, 2015: 134), a 'monolingual habitus' (Gogolin, 1993;, or a 'monolingual mindset' (Hajek and Slaughter, 2015), assuming languages to be bounded and capable of separation into teachable templates and able to be 'switched off' for most of the teaching week. As we have seen, the reality is more complex and heteroglossic than the IBDP policies suggest and, for many students, it is likely to be the case that, to use Evans' (1993: 164) words, 'the institutional structure is likely to be at variance with the map in the mind' of many multilingual students. This could have serious implications for students in that, as Cummins (2001: 19) reminds us, 'to reject the child's language in the school is to reject the child'. Multilingual students who language in ways different from the 'mother tongue' + 'foreign language' (the elite bilingual) model assumed by IBDP policies -those who cross imposed language boundaries, who mix named languages and cultures, and so on -might be disadvantaged and delegitimated by this monolingual habitus. As Apple (2013: 195) argues, 'the decision to define some groups' knowledge as the most legitimate, as official knowledge, while other groups' knowledge hardly sees the light of day, says something extremely important about who has power in society.' The seriousness of this risk of delegitimating multilingual students' knowledge and experience is more evident when we recognize that 'language is intricately linked to a person's identity and humanity' (Hopewell, 2017: 167). As Norton (2013: 45) has shown, 'it is through language that a person negotiates a sense of self within and across different sites at different points in time, and it is through language that a person gains access to -or is denied access to -powerful social networks that give learners the opportunity to speak'. In the IBDP, it could be argued that certain multilingual students have less opportunity to speak because policies play a role in prioritizing particular ways of languaging as more legitimate than others. As Ball (2006: 48) argues, 'we need to appreciate the way in which policy ensembles, collections of related policies, exercise power through a production of "truth" and "knowledge"'. For Ball, (2015: 307): teacher subjects and subject positions are formed and re-formed by policy and are '"invited" (summoned) to speak, listen, act, read, work, think, feel, behave and value' (Gee, Hull, and Lankshear, 1996:10) in particular and specific ways.
The way in which teachers and students speak, think, feel, and so on, about languages is shaped, at least to some extent, by the IBDP policies. It is in this way that 'the effect of policy is primarily discursive, it changes possibilities we have for thinking otherwise' (Ball, 2006: 49). Or, as Trowler (2003: 132) puts it, policies 'work to exclude other possible ways of conceptualizing the nature of education'. Of course, these polices (and a range of other material and discursive forces) influence multilingual students to varying degrees in different IBDP schools, but it is important to consider the extent to which the IBDP policies, in not having fully undergone a multilingual turn, might limit students' ability to realize the power of their multilingual identities and imagine alternative ways of learning and being. For multilingual students, the educational knowledge code underpinning the IBDP policies might impact upon their fundamental right to access their language repertoire, not just as a one-off scaffolding technique or a tolerated approach: but as an acknowledgement of their plurilingual identity, as a contribution to their emotional well-being and as recognition of children's agency in choosing how they prefer to learn a new language. (Ibrahim, 2019: 27) Is there cause for optimism that this situation might change -that the discursive power of IBDP policies could be used to legitimate multilingualism? It is important at this point to acknowledge the pragmatic barriers that might make it challenging for IB to implement curriculum, pedagogy, and assessment changes such as those outlined above that arise from the multilingual turn. However, the IB is a pioneering organization that has always sought to 'maintai[n] a balance between vision and pragmatism, always remaining true to its founding, visionary impulse, while dealing efficiently with the administrative work needed to run a global organization.' (IB, 2018a). The IB has innovated in the past in the interest of multilingual learners as it has sought to seek such a balance, for example by creating the (now discontinued) Language A2 course (Carder, 2007) for students who did not fit in the A/B dichotomy, and by producing documents such as Language and Learning in IB Programmes (IB, 2011) that engage with some of the implications of multilingualism. As their target market shifts more towards students learning in Type C international schools (largely 'host country nationals': Hayden and Thompson, 2013) with different language profiles to those in 'traditional' Type A international schools (largely for globally-mobile expatriates: Hayden and Thompson, 2013), IB has an opportunity to consider how their policies could be revised to meet the needs of multilingual learners in a way that balances pragmatism with a vision rooted in the multilingual turn. The IB dual language bilingual Diploma -in which students study their six subjects through two languages -long-established as the Gemischtsprachige IBDP in Germany (IB, 2017a), is now being trialed in Japan (Shrimpton and Ramos, 2018: 312-314) and soon will be in Korea (IB, 2019b). Perhaps this is a sign that future IB policies might undergo a more comprehensive multilingual turn.

Funding
The author received no financial support for the research, authorship, and/or publication of this article.