Anglonormativity in Norwegian language education policy and in the educational trajectories of immigrant adolescents

In the Nordic countries, policy debates about English often highlight the threat of domain loss for national languages, but the high status of English may also have a differential impact on people in Nordic societies. This article investigates a policy gap in Norwegian upper secondary education, whereby an advanced English subject requirement may hinder graduation for immigrant adolescents with little previous English instruction, despite English not being the medium of education in Norway. The aim of the study is to examine the impact of the upper secondary English requirement and of sheltered instruction as a local policy solution for such students. I use nexus analysis (Hult, 2015) to analyze ethnographic data from one upper secondary school that created an ad hoc sheltered English class. Data include field notes, classroom video and audio recordings, language portraits, and interviews with one school leader, one teacher, and six students. I draw on decolonial theory (e.g., García et al., 2021; Santos, 2007), notably Anglonormativity (McKinney, 2017), to trace discursive, interactional, and personal policy scales. I found that the sheltered class reflected discourses of integration and Anglonormativity, but nonetheless offered greater affordances for participation than a mainstream English class. Furthermore, comparing the emphasis on English remediation with students’ broader repertoires surfaced possibilities for reframing students as resourceful multilingual learners. I discuss policy options that might better address underlying issues of epistemic justice, compared to solutions limited to increasing students’ proficiency in languages of power like English.


Introduction
English occupies a high, if also contested, status in Norway, as in many places. In the Nordic countries, policy debates about English often concern the threat of domain loss for national languages in higher education, certain industries, and online (e.g., Källkvist & Hult, 2016;Linn, 2016;Phillipson, 2017). A more recent line of inquiry considers implications of the status of English for immigrants, whose experience with English can range from no formal instruction to fluency. Previous Nordic research has focused on the need for multilingual and multicultural pedagogical approaches to English in linguistically diverse classes (e.g., Burner & Carlsen, 2022;Tholin, 2014) and for including English in education for adult refugees (Dahl et al., 2018;Krulatz & Dahl, 2021). This article makes a new contribution by investigating a policy gap that arises in this context, whereby an advanced English subject requirement may pose a barrier to graduation for immigrant adolescents with little previous English instruction, despite English not being the primary medium of education in Norway. The aim of the study is to examine the impact of the upper secondary English requirement and of sheltered instruction as a local policy solution for such students, based on ethnographic data from one upper secondary school that created an ad hoc sheltered English class. I draw on decolonial perspectives (e.g., García et al., 2021;Santos, 2007), notably including Anglonormativity (McKinney, 2017), and nexus analysis (Hult, 2015;Scollon & Scollon, 2004) to trace the confluence of discursive, interactional, and personal factors in this policy dilemma.

English as a transnational school subject
English has achieved an unparalleled position as an additional language taught in much of the world, but conditions for English teaching vary greatly within and across countries, also in Europe (Gerhards, 2014;Lin, 2005;Phillipson, 2017). Norway is generally considered to have a high level of English proficiency (Linn, 2016). Although traditionally considered a foreign language, English increasingly operates as a second language in Norwegian society due to widespread use in personal, extracurricular, educational, and professional domains (Krulatz & Dahl, 2021;Linn, 2016;Rindal, 2014). For instance, the localization of English can be seen in shifting standards for English pronunciation, where a 'neutral' Norwegian accent has emerged among Norwegian teenagers as a desirable alternative to either British or American 'native speaker' norms (Rindal, 2014). Similarly, in Sweden, policy discourses frame English as a transcultural language, at once global and local, occupying a position just below Swedish in the country's linguistic hierarchy (Hult, 2012). This situation contrasts with a pronounced, if ambivalent, desire for colonially infused native speaker models in various postcolonial contexts (Lin, 2005;Lin & Motha, 2021) and in European contexts with a strong for-profit English teaching industry (e.g., Villacañas de Castro et al.,

3
Anglonormativity in Norwegian language education policy… 2018). Thus, the specificity of English in Nordic countries like Norway relates not only to its high status, but to a discourse of relative confidence in English as it is naturalized by Nordic users (see Hult, 2012;Rindal, 2014;Tholin, 2014). If English constitutes transnational linguistic capital (Gerhards, 2014;Lin, 2005) with high currency in Norway, immigrant students may find that this is capital they bring to their new school setting to varying degrees.
To the best of my knowledge, previous research has only briefly noted the potential for English to function as a barrier to graduation for immigrant students outside of countries where English is a primary medium of education (Beiler, 2021;Tholin, 2014). However, the situation has recently gained some public awareness in Norway, as reflected in public comments on a recent white paper, which proposes an introductory course in English for immigrant students deemed to need this (Official Norwegian Reports, 2019). In addition, a group of English teachers have written an open letter that calls for an alternative curriculum in basic English, modeled on the adapted curriculum in Norwegian that students can follow for their first six years in Norway (Gowie-Fleischer et al., 2021). Finally, the Ministry of Education and Research (2020) has noted another possibility: that other languages might be just as valuable as English in certain-notably blue-collar-lines of work, such that they might consider substituting a different language or subject requirement, at least for adult students.
The relevance of English in Norway's official introductory program for adult refugees has been considered in two linked studies (Dahl et al., 2018;Krulatz & Dahl, 2021). The authors found little emphasis on English in official mandates or educator conceptions of the program, whereas English was assigned greater importance in survey responses by refugees and in language requirements for higher education and selected job postings. Thus, they argued for greater emphasis on English in the introductory program. At the upper secondary level, Thomas and Breidlid (2015) have examined the rationale for and impact of high-stakes English tests in Norway. Based on critical discourse analysis of policy documents and supplemental interviews with English teachers and students, they argued that an emphasis on English primarily serves the interests of neoliberal globalization, while disadvantaging students from non-Western backgrounds. The latter claim is based on low reported levels of engagement with English in their student interviews and on overrepresentation of students with a non-Western background in the lowest bracket of standardized test scores. In response, Thomas and Breidlid (2015) called for reducing testing of English and instead allocating funds toward strengthening Norwegian proficiency among students with an immigrant background.
While differing as to the role of English, these studies share an emphasis on promoting immigrants' proficiency in a societally powerful language, whether this be English (Dahl et al., 2018;Krulatz & Dahl, 2021) or Norwegian (Thomas & Breidlid, 2015). This policy emphasis is understandable, given the high stakes for students. However, applied linguists working from a decolonial perspective have questioned the efficacy of simply increasing racialized and linguistically minoritized students' proficiency in majoritized languages (e.g., García et al., 2021;Motha, 2014;Souza, 2019). Instead, a decolonial approach involves addressing underlying epistemic inequalities in language education policy, as I will elaborate on in the next section.

Decolonial theory and Anglonormativity
Within decolonial theory, I include specific perspectives such as (de)coloniality (Mignolo, 2000), epistemologies of the South (Santos, 2007), and Southern theory (Connell, 2007). Central to all three perspectives is the assertion that knowledge production continues to reflect the epistemic divide between the colonizer and colonized that was necessary for the maintenance of European colonialism. On the side of the colonizer is modernity; on the side of the colonized, coloniality; and each can be seen as necessary for producing its opposite (Mignolo, 2000). Santos (2007) refers to this divide as the abyssal line, which separates knowledge from that which is not considered knowledge, the latter thus being relegated to the abyss of nonexistence. Similarly, according to Connell (2007), purportedly universal theory is produced on the 'modern' side of the line, while the colonial side is seen either as simply providing data or as creating particularistic, locally applicable knowledge. These epistemic inequalities are linked to the ascribed humanity of the knowledge producers, as the abyssal line also divides the fully human from the subhuman, seen not only in the treatment of people indigenous to colonized territories (e.g., in the Global South or Arctic North) but also of colonial 'others' in the North, such as refugees and undocumented migrants (Santos, 2007). Therefore, social justice must be premised on epistemic justice, "based on the recognition of the plurality of knowledges" (Santos, 2007, p. 64). This is a different critique of modernity than that offered within postmodern scholarship, which decolonial theorists see as continuing to center Western knowledge (Connell, 2007;Santos, 2007). Thus, Mignolo (2000) calls on scholars "to think otherwise, to move toward 'an other logic'-in sum, to change the terms, not just the content of the conversation" (pp. 69-70). Such epistemic decolonization is also needed in Nordic scholarship, which has, like other European scholarship, grown up under the tacit influence of the modernity-coloniality binary, manifested historically in colonization and assimilation policies especially in the Sámi north, but seen also in current exclusionary discourses toward non-Western immigrants (Eriksen & Svendsen, 2020;Gullestad, 2006).
Within language education research, García et al. (2021) have called for resisting 'abyssal thinking' (Santos, 2007) in studies that pertain to racialized bilinguals, that is, bilinguals who have been positioned as colonial 'others' by the abyssal line. Such abyssal thinking takes at least two forms: first, approaching racialized bilingual or multilingual students primarily in terms of remediation; second, overdetermining students to fixed types of linguistic personhood (García et al., 2021;Motha, 2014). The first of these commits epistemic injustice by consigning students' existing knowledge and language practices to the abyss of academic irrelevance, and both forms fail to recognize students' resources as complex and heterogeneous on their own terms. For instance, Chinese-English bilinguals may be erroneously assumed to need English language support, even if this is their primary language of communication, and to be more inclined toward studying math and science than a classical language like Latin (García et al., 2021). Similarly, Leung and Valdés (2019) have contrasted remedial language offers targeted at linguistically minoritized students with enrichment-oriented additional language instruction conceived for 'mainstream' students.
English has specifically been identified as a language implicated in hegemonic claims to knowledge and advancement, at the expense of other languages and forms of communication (e.g., Lin, 2005;Ngũgĩ, 1986;Phillipson, 1992;Souza, 2019). In post-apartheid South Africa, McKinney (2017) has recently developed the concept of Anglonormativity to describe the impact of the totalizing claims of English on students. According to McKinney (2017), "Anglonormativity then refers to the expectation that people will be and should be proficient in English, and are deficient, even deviant, if they are not" (p. 80). This norm is often linked to monolingualism and a racially white repertoire of English, notably in examples from South Africa and North America, although these linkages do not hold universally. For instance, in Hong Kong, Lin (2005) identifies elite Chinese-English bilingualism as the hegemonic norm, contrasted with subaltern, parochial Chinese monolingualism. Nonetheless, in all these cases, colonial varieties of English are prioritized. When such prestige varieties of English seem out of reach, students may simply retreat from classroom participation (Lin, 2005;McKinney, 2017). Although Norway differs in important ways from these settings, the discourse of Anglonormativity circulates globally (McKinney, 2017). Indeed, the global reach of English allows for considering the ideologically normative dimensions of English proficiency outside of former British colonies (Phillipson, 2017;Thomas & Breidlid, 2015;Villacañas de Castro et al., 2018). In using Anglonormativity as a sensitizing concept, I am also taking up the call for theoretical concerns that originate in the Global South to influence theory development in the Global North, thus challenging the privileged position of the North as the origin of transferable theory (Connell, 2007;Lin, 2005). I will argue that Anglonormativity sheds light on the unequal valuation of different languages in the Norwegian school system, when uncoupled from the normative nature of English monolingualism. In addition, the concept helps to account for the salience of English in the educational trajectories of students who have migrated transnationally, across settings where English might have occupied varying political and ideological positions.

Nexus analysis
I have used nexus analysis (Hult, 2015;Scollon & Scollon, 2004) to trace how discourses such as Anglonormativity are negotiated by educators and by students with transnational histories. Nexus analysis is an ethnographic approach to discourse analysis, which seeks to locate actions in space and time through mapping cycles of discourse as they come together at a nexus of practice (Hult, 2015;Scollon & Scollon, 2004). According to Scollon and Scollon (2004), this nexus is mediated by three dimensions: discourses in place (e.g., Gee, 1999), the interaction order (social norms as established in interaction; see Goffman, 1983), and historical bodies (embodied personal experience; see Nishida, 1958). Within language policy and 1 3 planning, important discourses in place include material and institutional arrangements, as well as written and unwritten policies that participants orient toward (Hult, 2015). Nexus analysis naturally directs attention to language policy arbiters, including educators (Hult, 2015). However, historical bodies can also include those who have borne the impact of language policy (e.g., Lane, 2010).
In this article, I extend the dimension of the historical body by critically engaging with theories of decoloniality, particularly those concerning knowledge production as embodied and, as such, socio-historically located (see Mignolo, 2000;Souza, 2019). In the concept of the historical body, nexus analysis relies on the Japanese philosopher Nishida's (1958) historicized and embodied understanding of how discourses and socialization trajectories become part of social actors' ways of being (Scollon & Scollon, 2004). This understanding aligns well with the emphasis on the body as a site of knowledge and epistemic inequality in decolonial perspectives (Mignolo, 2000;Santos, 2007;Souza, 2019). Furthermore, the focus on the body encompasses the researcher in both nexus analysis, where the researcher engages, navigates, and ultimately changes the nexus of practice (Scollon & Scollon, 2004), and in decolonial theories, notably in highlighting the locus of enunciation as a historical, geographic, bodily, and ideological position from which knowledge is produced (Mignolo, 2000;Souza, 2019). Indeed, it is particularly incumbent upon Northern scholars such as myself to make explicit my locus of enunciation, in order to acknowledge the partiality of the knowledge I may produce (Figueiredo & Martinez, 2021;Lin, 2005;Souza, 2019), as I do below.

Study context and aim
Two dimensions of Norwegian educational policy provide important context for the study: curricular language requirements and policies on organizational differentiation. Norwegian curricular language requirements broadly align with the Council of Europe's (e.g., 2022) recommendation for students to learn two languages beyond their first language. However, the requirements are structured according to the status of languages in society, more so than in individual students' lives. Norwegian and English are compulsory subjects during the first 11 years of school, and there is only flexibility in fulfilling any requirements for a third language. Notably, at the upper secondary level, students in general academic studies must pass a third, so-called 'foreign', language subject (most often Spanish, German, or French). However, this can be substituted by a self-study exam in any one of 43 languages, ranging alphabetically from Albanian to Vietnamese, or by study of an Indigenous (i.e., Sámi) or officially recognized minority language (e.g., Kven) in school (Norwegian Directorate for Education and Training [NDET], 2021).
English is not included in the curriculum governing foreign languages but is instead considered a separate subject, a division that reflects the status of English as a core school subject. Indeed, the national curriculum that was in force at the time of data collection (2017) portrays English as the key to international contact: "English is a universal language. When we meet people from other countries, at home or abroad, we need English for communication" (NDET, 2013, p. 1). This curriculum culminates in a first-year upper secondary course, with advanced competence aims such as the following: "understand and use an extensive general vocabulary and an academic vocabulary related to one's education programme" and "discuss and elaborate on culture and social conditions in several English-speaking countries" (NDET, 2013, pp. 10-11). These aims are illustrative of the standards that students must pass in English to receive an upper secondary diploma, which remain similar in the current English curriculum (NDET, 2019). Unlike in the Norwegian subject, there is no alternative curriculum for students who have not previously studied the language.
Another important characteristic of Norwegian educational policy is that organizational differentiation by achievement level is limited by law and relatively uncommon (Ministry of Education & Research, 1998). Nonetheless, some municipalities organize a separate upper secondary track for recent immigrants, called 'general academic studies for minority language speakers,' with separate classes in core subjects like math and Norwegian. The current study took place at a linguistically and culturally diverse upper secondary school in Southeastern Norway that had transitioned from a separate academic track to a mainstreaming model 3 years earlier. Concomitantly, five spots in each class were reserved for recent immigrants, who were exempted from competitive admissions. These students were referred to as "minority language speakers" by educators at the school. The school was prepared to provide students with supplemental Norwegian instruction and received funding to do so. However, they discovered that about half of the students admitted by exemption had difficulty passing the first-year compulsory English course, for which there was no official provision for additional teaching hours. The solution that the administration developed locally, despite lack of additional funds, was to offer a sheltered English class in which these students could repeat the first-year course before receiving a grade and potentially taking the national exam in English. The aim of this study is to examine the impact of the upper secondary English requirement and of sheltered instruction as a local policy solution for immigrant students with significantly less previous English instruction than their grade-level peers in Norway.

Methods, analysis, and locus of enunciation
The first stage of nexus analysis involves engaging the nexus of practice, notably in a situation of perceived inequality (Scollon & Scollon, 2004). I conducted 4 months of fieldwork at the participating school, which I chose due to several teachers' stated positive orientations toward their multilingual student body. I observed one mainstream and one sheltered English class taught by the same teacher (Lars 1 ) for 2 to 3 days per week, writing ethnographic field notes and making video and audio recordings of classroom instruction (Hult, 2015).
Students could participate in all or only some of these forms of observation (see Table 1). In addition, I interviewed the language department head (Astrid), the English teacher (Lars), and the six students in the sheltered English class. These students came from Iran, Poland, Russia, Syria, and Turkey. The five recent immigrants in the mainstream class were from Ethiopia, Latvia, the Netherlands, Poland, and Thailand. Eight of these 11 students created language portraits (Busch, 2012), which they then described. Both Astrid and Lars identified Norwegian as their only first language, and neither had an immigrant background. The interviews were conducted primarily in Norwegian, according to participant preference. Students could also record responses in other languages, which I then had translated. Data and participants are summarized in Table 1.
While the study topic emerged from ethnographic observation, interviews serve as a particularly prominent data source for two reasons. First, the interviews provided the most direct evidence of policy discourses framing the sheltered class. Second, the interviews made explicit comparisons between the sheltered and mainstream classes that were difficult to ascertain through observation alone. The classroom recordings provided supporting evidence of instructional practices (Hult, 2015), while the language portraits allowed students to represent their linguistic identities as lived and embodied (Busch, 2012).
The second stage of nexus analysis involves navigating the nexus by analyzing which scales contribute to bringing about social action (Scollon & Scollon, 2004). This stage began during fieldwork and continued in subsequent analysis, during which I employed a recursive and comparative process to identify connections and tensions among data sources (Hult, 2015). To gain a comprehensive overview of themes, I first coded all field notes and interview transcripts with eclectic codes, which I then consolidated through focused coding (Saldaña, 2015). Through this second coding cycle, I identified discourses, interactional patterns, and biographical experiences that elucidated sheltered English teaching as a nexus of practice (Hult, 2015;Scollon & Scollon, 2004), leading to a focus on integration and Anglonormativity (discourses in place), classroom participation (the interaction order), and trajectories of emergent multilingualism (historical bodies). I extended and nuanced my understanding of these themes by reviewing the classroom recordings and students' language portraits, especially as these provided insight into the interaction order and students' historical bodies, respectively. In line with a goal to recognize different forms and sources of knowledge (Connell, 2007;Santos, 2007), I attempt to highlight the voices of educators and, especially, students who live with the impact of the policy in focus. Still, this study represents my own account of events. I do not believe I can 'give voice' to others, especially racialized or linguistically minoritized students (Heller, 2008), even as I seek to describe their emic perspectives as best I can from my personal and social position. My gaze is influenced by my experiences as an English teacher educator raised bilingual in English and Norwegian, identified as white in North America but sometimes perceived as foreign in Norway. I further acknowledge the significance of geopolitical location for theory development, such that I seek to offer insight into the localization of English in a Nordic context, rather aiming for a universalizing description of the global dynamics of English (see Hult, 2012;Figueiredo & Martinez, 2021;Souza, 2019). Finally, through this analysis, I hope to contribute to changing the nexus of practice (see Scollon & Scollon, 2004) in ways to which I will return in the discussion.

Discourses in place: integration and Anglonormativity
As there was no official policy that called for sheltered English teaching, salient discourses in place could most easily be traced through participant interviews, especially with the language department head, Astrid. Two policy discourses consistently framed the decision to create the sheltered English class. Of these, integration was the more explicit discourse, which the educators appealed to even as they explained the decision to create a separate English class for certain immigrant students 2 (see Extract 1).

Extract 1 (interview)
Ingrid: does that mean that these like are solutions that you have come up with locally […] but you don't receive any funding for it? it's just-Astrid: we have to take it from the budget the school has yeah, mm (0.9) and before we had separate linguistic minority classes, three-year program, in general academic studies (0.5) but we have gone away from that because we shouldn't have like, organized (1.4) mm (0.7) segregation with differentiation or, it should happen within the class, it should be integrated and they probably learn more from that, in most cases, but if the level is as low as it is for some of them in English (0.9) then it just becomes awful to sit in the English classes and (1.0) extinguishes completely, no mastery, no experience of participation, nothing, while they can manage fairly well in other subjects in that class, while (1.4) yeah, one can image how it can be to sit in an Arabic class and like know five words too, while the others discuss famous authors and films and societal issues in the target language In Extract 1, Astrid explains that the school leadership has carved out the funds for the sheltered class from the school's regular budget. She traces this development to the transition away from a separate program for recent immigrants. Astrid draws upon a widely circulating discourse in Norwegian educational policy, that there should be integration and differentiation within the class, signaling with a long pause the taboo nature of the alternative, "segregation with differentiation." Thus, she aligns with mainstreaming as a preferable policy and as the organizational realization of integration. However, Astrid also introduces a competing consideration: in some cases, students know too little to learn in an 'integrated' setting. This situation leads not to participation, but to students shutting down, which she captures with the metaphor of a light being extinguished. This line of argument relativizes the discourse of integration by considering how it connects with the historical bodies of students to produce an undesirable interaction order. Astrid also demonstrates an ability to decenter by imagining a reverse scenario, in which Arabic, rather than English, is the target language, and Norwegian students must manage conversations about literature, film, and societal issues with a hyperbolically small vocabulary of five words in Arabic. Thus, she recognizes the arbitrariness with which discourses position particular historical bodies as linguistically deficient.
While integration comprises a background discourse for decisions about instructional organization, the more direct impetus for creating the sheltered class is the prominent status of English in the Norwegian school system (see Extract 2).
Extract 2 (interview) Astrid: they automatically get Norwegian, and English can vary from year to year, but for the most part, there has been a need for it, because far from all countries have as much English instruction as what our students have, so, why have we chosen to do it that way? we would of course like the students to pass and get through and (0.5) manage to get a good life, and reach-reach some dreams they might have, and if we don't help them with that, then they don't get a diploma (0.5) we can of course say that yes it's good for our statistics, but that's not what's important, the important thing is that (0.5) the students get-(1.4) manage to get a diploma (0.8) mm, so that's a driving force for us, yeah Ingrid: and how did you come up with that solution?
Astrid: well I guess it was just that we saw the need, it was necessary to do something […] we have (0.7) just seen that (1.0) complete integration, with just attending regular classes and not receiving anything extra, it-it just doesn't work, then it became a worse offer than what they had in-when we had linguistic minority classes (0.6) and (0.8) that had its weaknesses because they got very few Norwegian impulses when they attended separate classes (0.8) but then there's the fact that they should master, they should manage, and if they sit there in English classes and don't understand anything, then that creates a kind of-(1.1) like what it does to the self-image of a 16-year-old, that's not good at all, they certainly struggle enough, so then that's just a-find a hole in the schedule because-(0.8) and then find a teacher who suits, get started, we must-that's necessary for them to have a chance to manage As Astrid describes in Extract 2, the official policy is clear and severe: without passing the compulsory English course, students will not receive an upper secondary diploma. This requirement positions some immigrant students' educational backgrounds as insufficient, since "far from all countries" include as much English in the curriculum as Norway. Astrid delineates two possible responses, one hypothetical and the other actual. She calls into play and then backgrounds an educational monitoring discourse, by stating that the school's statistics would have benefited from ignoring the problem, presumably since students who drop out would not contribute to lowering the school's grade point average. Instead, she foregrounds a concern with students' prospects, arguing that graduation is a premise for students to have a "good life" and "reach some dreams." This concern led her and her colleagues to "do something." The juxtaposition of these competing discourses is a reminder of the students' precarious situation, as other local policymakers could have resolved discursive tensions differently (Johnson, 2009).
At a more overarching level, two policy discourses frame Astrid's account in Extract 2. First, Anglonormativity (McKinney, 2017) operates at the institutional level as a negative judgment against students' educational backgrounds and qualifications for graduation, insofar as the students cannot demonstrate high English proficiency. In addition, Astrid suggests that the effect of an inability to participate in English class on "the self-image of a 16-year-old" is "not good at all," indicating the harm that Anglonormativity may cause when students internalize this discourse in their historical bodies (see also Lin, 2005). Second, Extract 2 echoes the discourse of integration (see Extract 1) and further supports a reading of integration as mainstreaming, as Astrid explains that a completely separate program did not provide enough "Norwegian impulses" to immigrant students. Thus, while Anglonormativity and integration push toward different organizational arrangements, both index immigrant students' subaltern position in the educational system.
The discourse of Anglonormativity also became salient in students' accounts of their transition to school in Norway. The students in the sheltered class all described curricular demands in English as challenging, due to the later start or different nature of English teaching in their previous home countries. For example, Sara noted that English teaching in Poland had been less communicatively oriented. Sara further described Anglonormativity in encounters with other, presumably Norwegian, students: Extract 3 (interview) not everyone understands that maybe in my home country, there isn't like the same-same level of English and such like there, ((taunting)) o::h you're like sixteen, don't know English, ((unmarked)) but they don't know, it's not like (0.5) I don't want to, but I just didn't have the opportunity to learn for real In this extract, Anglonormativity (McKinney, 2017) is evident in the assumption that all societies should have high English proficiency, with Norway upheld as an exemplar of the norm. However, unlike in settings where English monolingualism is dominant (e.g., García et al., 2021;McKinney, 2017),

3
Anglonormativity is further tied to age. Low English proficiency is positioned as childlike, as the fact that Sara is sixteen is invoked in the ventriloquism of an unsympathetic peer. This boundary-making positions majoritized Norwegian bodies as normal, by portraying English proficiency as something to be acquired by adolescence, though not necessarily in early childhood, a measure against which most Norwegians would compare unfavorably. Indeed, Anglonormativity becomes a discursive tool in reported peer interactions to constitute Sara's difference as an immigrant. Similar language policy discourses have been reported in Sweden, which position English and Swedish as intertwined elements of national identity, whereas immigrants' minoritized languages are relegated to the periphery of society (Hult, 2012;Tholin, 2014).

The interaction order: patterns of participation
The interaction order of classroom instruction may become especially evident when it differs across settings (Scollon & Scollon, 2004). A significant contrast between the mainstream versus the sheltered class found expression in patterns of participation. It was not possible to trace specific students' participation patterns across different types of English classes, given the duration of the study. However, comparison was possible in two ways: first, through observation of participation among current students in the sheltered class versus those in the mainstream class who could later take the sheltered course; second, through teacher and student narratives, especially since Lars had taught some of the same students first in a mainstream class and then in the sheltered class.
As noted above, there were five students in the mainstream class admitted through exemption from competitive admissions, based on their length of residence in Norway. As for others in the class, these students' participation varied greatly. George, who had moved to Norway from Poland a few years earlier, was outgoing and participated vocally in both small group and whole-class discussions in both English and Norwegian. Bob, a more recent immigrant from the Netherlands, was similarly vocal in English but less comfortable speaking Norwegian. In fact, he transferred early in the school year to an English-medium International Baccalaureate program. In contrast, I did not record any instances of Lemi, Geo, or Pandy, from Ethiopia, Thailand, and Latvia, respectively, participating vocally in whole-class discussion during my 4 months of observation. Lars considered recommending some of these final three to the next year's sheltered class. Of course, it is not possible to explain these students' less vocal participation simply in terms of educational or national background, as participation among students who were raised in Norway also varied. Instead, this pattern of relative silence primarily has analytical significance because it aligns with an interaction order described by both Lars and the students in the sheltered class.
Indeed, the students in the sheltered class all described withdrawing from interaction in their previous English class because of feeling too far behind their classmates. Esperanza, a student from Russia who participated very actively in the sheltered class, compared her experiences of her mainstream and sheltered English classes as follows: Extract 4 (interview) Esperanza: I had very little (0.6) knowl-knowledge in English so there were lots of students who knew more, and for example, some of them, uh are from the USA, so therefore I just (0.5) listened and I-I couldn't-(0.6) um (0.9) didn't have time to get information and (0.6) pro-process it in my head, while they already (0.9) answer so I just, oh, it was a bit of brainstorm 3 ((laughs)), yeah that was difficult, but now I know a little more and he tells-explains better for us, and therefore I have time to reflect and answer In Extract 4, Esperanza describes herself as having little knowledge in English compared to others in the mainstream class, underlining the epistemic impact of being constrained to this single language. In particular, she relates that other students would answer before she had time to process information. She experienced this as a "brainstorm," making creative use of this English idiom to denote a troubling state of disorder in her mind. She contrasts feeling passed by in the mainstream class with her current situation, where the teacher explains more thoroughly, and she has more time to reflect and answer, that is, to construct and express knowledge. Many of the historical bodies have remained the same, but the interaction order has changed. Esperanza and I co-construct this situation as one of "new possibilities." Another student, Ecem, echoed Esperanza's account of whole-class interaction and extended this withdrawal to interactions with students in the mainstream class (see Extract 5). In Extract 5, Ecem refutes my premise that she would work with other students in the mainstream English class. She initially explains her withdrawal as being due to the preponderance of Norwegians, whom she associates with high English proficiency. However, she adds that two other "foreigners" also knew English, even "very perfect English" in the case of an Iranian student. The latter is signaled as somewhat unusual by the conjunction "but" after Iran. Ecem, being herself Turkish, thus associates English more strongly with European than Middle Eastern bodies, though not without exception. She further sets up an opposition between herself and a current classmate from Syria "who didn't know English at all" versus an emphatic "they" or "everyone" who "knew English, and at a very high level." If other students tried to include her and her current classmate, they would refuse the initiative. Whereas Motha (2014) writes of teachers who excluded emergent bilinguals in the United States from classroom participation, Ecem, Esperanza, and other classmates described an interaction order whereby they themselves withdrew from participation in the mainstream setting (see also McKinney, 2017), despite encouragement to the contrary.
The teacher, Lars, also noticed a difference in levels of participation among students who had transitioned from the mainstream to the sheltered class. He said of the sheltered class: it is a great offer, because these students finally dare, it's the second year right, to participate actively in class and raise their hand and talk and participate and all that, like [a student] who is very vocally active, she barely said a word (0.6) in this general studies class she was in last year […] and that's because there she felt that here, this wasn't her level, here I don't dare to contribute much According to Lars, the sheltered class provides greater affordances for students to participate, illustrated through one student whom he taught in both mainstream and sheltered settings. Lars attributes students' previous withdrawal to intimidation at the advanced level of English in the mainstream class, an impression that this student confirmed separately (see also Extracts 4 and 5). Again, English proficiency has epistemic consequences, as discomfort with speaking English results in silencing of knowledge, or 'contributions,' in class.
My observations in the sheltered class confirmed the students' and teacher's accounts of greater affordances for participation in this setting, as in Extract 7. The sequence is part of a teacher-led discussion of the 'American dream' and the author Horatio Alger.
Extract 7 (classroom video) 1 Lars: so anyone, uh please explain to me, what does that, expression mean, to go from ra::gs, to riches (7.0) ((nervous smile, then points to)) Ecem 2 Ecem: ok, I can try, um this means start-start liv < life > with nothing but be-bec-I don't know um In the first turn of Extract 7, Lars poses a question and then waits several seconds for an answer. Given Ecem's reported resistance to participating in the mainstream class (Extract 5), it is significant that she volunteers (turn 2). When Ecem stops and indicates uncertainty, Lars encourages her, with a rising final tone, to keep trying (turns 3-7). Ecem then receives help from her Palestinian classmate Rachelle to pronounce the word become (turns 8-10), with which Ecem continues her answer (turn 12). The departure from English is met with support rather than sanction, even if recourse to languages other than Norwegian was rare (see Beiler, 2021). At this point, Lars seems to pick up more on Ecem's final expression of uncertainty than her substantial response, as he asks for another volunteer and gives the floor back to Rachelle (turns 13-14). Nonetheless, Ecem has held the floor for much of the sequence. This extract illustrates a broader pattern in the sheltered class, whereby students would often risk evaluation despite some uncertainty about vocabulary, pronunciation, or desired answers. Such affordances for participation contrasted markedly with reported experiences of the mainstream setting (see Extracts 4-6), even if students' knowledge only seemed to find partial expression in the sheltered setting as well. In sum, the interaction order in the sheltered class positioned these students as more legitimate participants, as the discourse of Anglonormativity operated more weakly than in the mainstream setting.

Historical bodies: trajectories of emergent multilingualism
The two previous sections arguably describe a positive, if limited, resolution to a language policy dilemma, in which local policy actors created an instructional arrangement that both teachers and students found helpful. However, a decolonial perspective pushes us to look beyond successful remedial education to analyze the reproduction of racialized and linguistically minoritized students' inferior positioning in educational systems (García et al., 2021;Motha, 2014). An analytical imperative that follows is to approach students not primarily in terms of what they lack, as English language learners, but what they possess, as emergent bilinguals or multilinguals (García, 2009). This third findings section therefore focuses on students' broader trajectories of emergent multilingualism and how learning English fit into these, analyzed through the lens of the historical body that carries experiences and practices into new social encounters (Scollon & Scollon, 2004).
To understand immigrant students' possibilities for negotiating the previously described discourses in place and interaction order, it is necessary to consider both embodied histories and desires (Lin & Motha, 2021;Scollon & Scollon, 2004). A student called Lamis illustrates well the limited linguistic subject positions available to immigrant students because of her evident difficulty with the English subject, even in the sheltered class. Lamis came to Norway as a refugee from Syria, after which she attended an introductory class for one year and then mainstream education for 4 years before the study. In Figure 1, Lamis has represented her embodied experience of language in the form of a language portrait (Busch, 2012).
In the top left corner of Figure 1, a box labeled "English" is connected to the ear, where Lamis has written in Norwegian, "[I] hear [it]" and "difficult to use". "English" also appears with a Wi-Fi icon, with which Lamis meant to indicate that she had a weak signal in English. In both cases, English is portrayed as external to her body. In contrast, four elements are internalized. On the raised arm, Lamis has drawn the French flag with the phrase "Je t'aime" ['I love you'], associating French with love and romance, as several other students did. Lamis commented that "French is my favorite language. I love French" (interview).
The Norwegian flag appears in the place of her mouth, as Norwegian is what she speaks in most contexts at this point. In the place of her heart, Lamis has drawn the colors of the Kurdish flag, symbolizing her close identification with Kurdish as her mother tongue. Finally, several Arabic letters appear on her head and torso. Lamis explained, "Arabic is everywhere, like the whole body, you see the head, what I have used […] it's not my favorite language" (interview). Here, Lamis carries in her body discourses from Syria, where Arabic was ubiquitous, including as the medium of education, whereas literacy instruction in Kurdish was illegal. This history of repression may have contributed to Lamis's ambivalent relationship to Arabic, which she used infrequently as a support in English class, compared to her Palestinian classmate Rachelle, who had also fled from Syria.
Despite a negative positioning within the discourse of Anglonormativity, Lamis narrated experiences of success as a language learner in French classes in Norway (see Extract 8). In Extract 8, Lamis describes French class as easy and relates that she received a grade of five out of six, considered a very good grade. In contrast, she told me earlier in the interview that she had barely passed English. Lamis describes the difference between the two language subjects in terms of content. In French, they "learned the language" first, including spelling and grammar, not French culture. This can be understood as a matter of emphasis and sequencing, as the French subject does include cultural content at all levels. However, in English, she states three times that she "didn't learn the language." Instead, she joined a course focused on culture and ways of life in English-dominant countries (e.g., Extract 7), which presupposed many years of previous study. Lamis was emphatic throughout her two interviews that she needed to learn basics of the English language before she could engage in discussions of English-language literature and culture. Thus, Lamis's account of English class resembled the hypothetical Arabic class Astrid described in Extract 2. Within the discourse cycles of nexus analysis, the difference between French and English classes in Norway can be understood in terms of assumed personal histories, in ways that treated Lamis's historical body as a legitimate participant in the interaction order of the former but not the latter.
Furthermore, the comparison between English and French challenges two broadly circulating discourses within language education and research. First, it disputes a typological explanation for Lamis's difficulty, as English and French are about equally different from the languages she learned in childhood (cf. De Angelis, 2007;Tholin, 2014). Second, it locates the problem in the curriculum, rather than in English or language learning as such (see also Tholin, 2014). Lamis managed to learn both Norwegian and French well, but in English there was a greater gap between the amount of instruction received and the expectation of performance. This comparison thus challenges a tendency to channel linguistically minoritized students toward linguistic remediation, typically in a majority language, and away from enrichment-oriented world language studies (see García et al., 2021;Leung & Valdés, 2019;Motha, 2014). Indeed, Lamis, along with several other recent immigrant students, displayed a significant desire to learn languages such as French, Spanish, and Japanese. Furthermore, most of these students also expressed strong investment in English, with justifications ranging from pragmatic resignation to extracurricular enjoyment (cf. Lin, 2005;Thomas & Breidlid, 2015).

Discussion and conclusion
The impact of the upper secondary English requirement on students participating in the sheltered class can best be understood through the dynamic convergence of discursive, interactional, and personal scales (see Hult, 2015). At the discursive level, a partial policy vacuum lent importance to local policymakers' negotiations of widely circulating discourses about language education and immigrant students. Notably, the discourse of integration as mainstreaming, strongly present in national educational policy (Ministry of Education, 1998), gave way to sheltered teaching due to the uptake of another powerful discourse circulating at a global scale, namely Anglonormativity (McKinney, 2017). This uptake by local policymakers did not result from an assessment of one discourse as more compelling than the other in abstract terms. Rather, the encounter of Anglonormativity with students' historical bodies led to an interaction order that rendered the appearance of both mainstreaming and learning impossible. In the face of such layered effects, school leaders' dispositions emerged as decisive, seen in their ability to critique the curricular expectation as unreasonable, such that they felt compelled to make alternate arrangements. Thus, the nexus of practice was significantly shaped by the historical bodies of local policymakers, as well as those of the students about whom policy decisions were made (cf. Källkvist & Hult, 2016;Lane, 2010). A contribution of this study to nexus analysis may be to critically consider not only how historical bodies act, but also how they are met, within the nexus of practice, notably through bringing nexus analysis into dialogue with epistemologies of the South (Santos, 2007) and decolonial perspectives (e.g., García et al., 2021;Mignolo, 2000).
This study has identified a gap between certain immigrant adolescents' historical bodies and an Anglonormative educational policy, yet multiple policy uptakes are possible. Existing policy recommendations emphasize the need for better adapted and more formalized compensatory English teaching (Dahl et al., 2018;Gowie-Fleischer et al., 2021;Krulatz & Dahl, 2021;Official Norwegian Reports, 2019). However, a decolonial perspective calls for first 'thinking otherwise' about a macro-narrative that frames immigrant students in deficit terms (García et al., 2021;Mignolo, 2000). Therefore, turning critical scrutiny toward Anglonormativity rather than toward students' repertoires, I would like to consider which resources and capabilities an Anglonormative policy fails to apprehend and which policy arrangements might better reflect these.
An Anglonormative policy might be said to relegate all but English to the abyss of educational insignificance (McKinney, 2017;Santos, 2007). However, in a context such as Norway, Anglonormativity is added onto a language regime that first prioritizes the national language (here, Norwegian; see Hult, 2012, for Swedish). The emphasis on Norwegian positions immigrant students as a whole as needing linguistic remediation to benefit from education (Burner & Carlsen, 2022;Official Norwegian Reports, 2019;Thomas & Breidlid, 2015), whereas Anglonormativity adds distinctions among students based significantly on the role of English in their previous education. These distinctions may privilege students from Northern Europe, the United Kingdom, and former British settler-colonies, though not absolutely so, as some counterexamples in this study suggest. Regardless, other resources are treated as relatively unimportant or at least as more interchangeable. Only English and the national language constitute core linguistic knowledge (see also Hult, 2012;Tholin, 2014).
A decolonial perspective suggests that the solution is not simply to help students adapt to this norm, but to imagine and work toward 'otherwise' arrangements (Mignolo, 2000). One seed for this work might be to challenge the overdetermination of immigrant students to remedial forms of language education, even bilingual ones (García et al., 2021). This could entail a more flexible multilingual education policy, where students could pursue other languages more fully than English, either languages that are new to them (e.g., for Lamis, French) or ones where formal study has previously been denied (e.g., for Lamis, Kurdish). Such arrangements would better acknowledge the complexity of students' capabilities. Moreover, a more flexible policy could resist the colonial push toward "a single homogenous future" (Santos, 2007, p. 50) tied to English by, instead, envisioning multiple successful outcomes to language education. Still, Lin and Motha (2021) caution that the hegemony of English persists not only through official policies but also through desires internalized by historical bodies. Indeed, changing the nexus of practice must go beyond top-down policy processes to engage various stakeholders in critical reflection about desirable linguistic subject positions (García et al., 2021;Lin & Motha, 2021).
Another seed could be to reconsider English as a fixed object and instead emphasize the broader educational purposes to which English may contribute. In the Brazilian secondary curriculum, Souza (2019) argues that English operates "more as a pedagogical device than as a natural language" (p. 24), thus subordinating English to local needs for developing understanding of and respect for difference. Such aims are also present in the Norwegian curriculum in English, as when it states that "English shall help the pupils to develop an intercultural understanding of different ways of living, ways of thinking and communication patterns" (NDET, 2019, p. 1). If we look across the abyssal line (Santos, 2007) to students' silenced knowledge and language resources, we might see that immigrant students often already possess transnational and translingual experience, networks, strategies, and resources far beyond those of students who develop elite bilingualism through formal schooling (see Canagarajah, 2013;García et al., 2021;Tholin, 2014). The fact that those experiences and resources count so little in formal language education can be justified on pragmatic grounds, but it must also be named as an epistemic injustice. A more just language education policy would give students greater opportunities for demonstrating such competences and expertise within and beyond the bounds of any given language subject, for instance through translingual practice (Canagarajah, 2013).
In conclusion, educational language policy needs to not only provide adequate support for meeting requirements in English, but moreover to apprehend the resources that minoritized students possess. A decolonial approach can contribute to such revisioning by drawing attention to multiple forms of expertise (Santos, 2007). Language education policy needs to recognize multiple pasts and facilitate multiple futures to a greater degree than today, without consigning minoritized and racialized students to future educational and vocational trajectories that are positioned as inferior (García et al., 2021;Lin & Motha, 2021). This means both seeing immigrant students as capable of learning languages of power, when provided with time and opportunity, and raising the status of the resources and experiences they already possess. Perhaps most important is an explicit recognition that educational language requirements do not simply reflect 'real-life' demands. To the extent that they uphold English and powerful national languages like Norwegian as the primary resources for success, they also reinforce the abyssal line between recognized and silenced forms of knowledge (Santos, 2007).