Students navigating shifting literacy expectations in grade 6 history teaching: a qualitative text-based study

Abstract This qualitative study focuses on Grade 6 students’ possibilities to engage in disciplinary writing practices in history teaching in a school located in a socially disadvantaged and linguistically diverse area. The aim is to contribute knowledge about how students negotiate different literacy expectations in the teaching. The researcher used ethnographic methods to follow one teacher and two groups of Grade 6 students for 12 weeks’ teaching about the Vasa era in Swedish history. The material used consists of samples of students writing, parts of the text material they studied, and written questions they answered in tests and pre-writing discussion. In the analysis, systemic-functional linguistics (SFL) and thematic content analysis were employed to develop categories of literacy expectations based on the written questions. SFL was also used to analyze samples of the students’ writing and relevant textbook material. The findings show that the expectations increased in a final test because most questions required causal reasoning. In contrast, the questions used in text discussions and homework tests conveyed a variety of literacy expectations, with an emphasis on chronological reconstructions of events. While some students adjusted their writing to causal reasoning, others relied on a reconstruct approach that adhered closely to the simple text material that was chronologically structured. Outlier examples showed students using multimodal representations, a substitute teacher’s oral elaboration on a text, and resources of learner language in their answers. Implications for teaching that support all students’ engagement in disciplinary practices of history are discussed.


Robert Walldén (PhD) is Associate Professor of
Swedish and Didactics at Malmö University, Sweden. His research interests include disciplinary literacies, text practices, and knowledgebuilding interaction in linguistically diverse classrooms. His research often builds on social semiotics, Bernstein's sociology of education and critical discourse analysis. An important aim is to shed light on possibilities for promoting students' awareness of texts and language in ways which enables them to move between different disciplinary discourses and looking at them critically. The research is conducted within the research program Disciplinary literacy and inclusive teaching (LIT) at the Faculty of Education and Society at Malmö University. LIT researchers explore the multidimensional nature of language in teaching and learning processes, aiming to gain a better understanding of the development of literacy in different educational levels and settings in which teachers, students, and learning materials interact.

PUBLIC INTEREST STATEMENT
Across educational contexts, there is a shared concern about young students' development of literacy skills necessary for school achievement and democratic participation in society. The educational challenge is particularly salient in classrooms in which many students are taught in their second or additional language. Conducted in two linguistically diverse Grade 6 classrooms in Sweden, the present study highlights students' disciplinary writing in the context of primary school history teaching. Through a novel qualitative approach, the study explores the different demands put on the students writing and how the students deal with these demands. The study shows the importance of students being sufficiently scaffolded to produce the complex writing required in summative test situations. This scaffolding requires increased attention to writing in the classroom, selection of highqualitative text material to learn from and giving students rich opportunities to negotiate disciplinary content multimodally.

Introduction
In recent years, there has been a growing research interest in students' opportunities to develop disciplinary literacies through reading and writing. Mirroring developments in other countries, for example, the United States (Wissinger et al., 2018) and Finland (Rantala & Khawaja, 2018), the Swedish national curriculum since 2011 has required teaching to have a more disciplinary focus. This has impacted the teaching of social studies subjects (civics, geography, history, religion) in elementary schools since they are taught and assessed individually rather than as a unit from Grade 4 (e.g., Samuelsson, 2014). The shift to disciplinarity has occurred in tandem with increased linguistic diversity and school segregation with respect to students' migration backgrounds (e.g., Bunar & Ambrose, 2018;Hansen & Gustafsson, 2016). This has brought further attention to the linguistic dimensions of subject-specific reading and writing practices to support students' receiving instruction in the majority language, Swedish (e.g., Nygård Larsson, 2011;Olvegård, 2014;Uddling, 2021). International research on disciplinary teaching and learning in linguistically diverse settings stresses the importance of a language-conscious mode of teaching (Iddings, 2021;Llinares & McCabe, 2020;Moore et al., 2018). The present study will focus on test-related writing practices in history teaching in Grade 6 in a school located in a socially disadvantaged and linguistically diverse area.
As the present section will show, texts and writing in history teaching have been substantially researched (see also Theoretical perspectives). Among different historical genres, previous research has particularly concentrated on the writing of argumentative texts, a genre generally perceived a challenging (e.g., Martin & Rose, 2008;McKeown et al., 2019). Furthermore, argumentative writing is often associated with sourcing skills, which is considered a crucial and distinguishing part of historical disciplinary literacy (e.g., Seixas & Morton, 2013;Shanahan & Shanahan, 2008;Wineburg, 1991). Intervention studies in Grade 6 (Wissinger et al., 2018) and Grade 8 (De La Paz et al., 2017) have shown that students' disciplinary writing and learning improve with the teaching of cognitive strategies that involve learning sourcing skills and conventions for structuring historical arguments.
The prevalent research focus on written argumentation based on sources in history (see also, Fitzgerald, 2019;Rantala & Khawaja, 2018;Wissinger et al., 2018) entails that other genres frequently used in teaching the subject, such as explanations and recounts, have received less attention. Olvegård's (2014) study of Swedish upper-secondary history textbooks found that arguments were almost absent in the material. Moreover, explanations were rare compared to accounts and descriptions. A study of South African history textbooks (Bharath & Bertran, 2015) showed that argumentative writing was absent until Grade 9, while Grade 6 textbooks transitioned from recording genres (e.g., recounts) to explaining genre. While the Swedish curriculum implemented in 2011 (Skolverket, 2019) placed demands on students' disciplinary thinking and capabilities to engage in argumentation and explanation in social studies teaching, researchers found these literacy expectations to be implicitly phrased and mostly evident in summative assessment situations, such as the administration of national tests (e.g., Molin & Grubbström, 2013;Samuelsson & Wendell, 2016;Staf, 2019). Thus, researchers have argued that students are not sufficiently prepared in teaching to meet these expectations.
A revised Swedish national curriculum implemented in Fall 2022 (Skolverket, 2022) marks a partial return to a more "factual" content area teaching and assessment approach, particularly in the early grades (1-6). As such, the curriculum is unlikely to counteract students' tendencies to read textbook material in a factual (rather than critical) way (e.g., Rantala & Khawaja, 2018;Wineburg, 1991). In the new history curriculum, interpretation of sources is not expected in Grade 4-6 (cf., Samuelsson & Wendell, 2016) but only in 7-9. Other literacy expectations will remain but are even less explicit than in the former curriculum. Notably, the 2011 curriculum had a separate "core content" section for Grads 4-6 with historical concepts, including change, chronology, cause, consequence, and interpretation (Skolverket, 2019). Although this explicit attention to metalinguistic and historical concepts is absent in the revised curriculum, the students are still expected to learn and communicate knowledge about "causes and consequences of societal changes and of people's actions" (Skolverket, 2022). The concern of students and teachers navigating such ambiguous literacy expectations lies at the core of the present research.
According to the nuanced understanding of literacy which have evolved throughout the last decades, writing is no longer considered an isolated skill but part of social practices that also involve speaking, reading and multimodal forms of expressions (e.g., Bezemer & Kress, 2016;Ivanić, 2004;Street, 1984). In history, writing is often based on the reading of textbooks and historical sources (Coffin, 1997;Monte-Sano et al., 2017;Shanahan & Shanahan, 2008). However, classroom studies that highlight the students' writing in relation to the text material they learn from tend to study genre-based instruction (Rose & Martin, 2012), in which the joint deconstruction of texts is integral to the scaffolding writing process (Fenwick & Herrington, 2021;Hallesson & Visén, 2018;Harman, 2013). As also shown in other studies on text practices in history teaching (Kindenberg, 2021;Matruglio, 2021;Schleppegrell et al., 2004), this places the focus on the teacher's knowledge and administration of a specific pedagogic metalanguage. In the present study, I take a different approach by connecting Grade 6 students' test-related writing to the text material they have studied. I will trace literacy expectations across different questions posed to the students in test situations and pre-writing discussions and put these questions in relation to both the students' writing and the characteristics of the texts they have studied to support their learning. This enables me to highlight writing practices and literacy expectations in history teaching in a way that has not been achieved in previous research. The purpose of the study is to contribute knowledge about how students negotiate literacy expectations when writing answers to test questions in the teaching of history. The research questions are as follows: (1) Which literacy expectations are expressed in written questions administered to the students in a final test and in activities leading up to the test?
(2) How can the students' written answers be understood in relation to the literacy expectations in the written questions?
(3) How can the students' written answers be understood in relation to properties in the text material they studied in preparation of the tests?
The study builds on previous publications partly based on the same material that focused on notetaking and pre-writing activities. Walldén and Lindh (2021) showed how the participant teacher guided the students' notetaking for purposes of storing information in the text material and participating in open discussion about the content. This involved the students using multimodal representations, such as mind maps, and expressing thoughts about content-related images in discussions and writing. Walldén (2020b) focused on the peer interaction and textbooks that related to the teaching. An important finding was that the simplified nature of the texts seemed to limit students' possibilities to engage with disciplinary language and content. The present article focuses on the students' test-related writing and written questions administered to the students in written tests and pre-writing discussions.
In the next section, I will describe the linguistic perspectives underpinning the analysis of the text material. While such perspectives are often used to distinguish between "strong" and "weak" writers (e.g., Myhill, 2008), it is important to go beyond the analysis of individual texts and consider the wider context. In Nordic countries, classroom researchers have discussed an insufficient attention to metalinguistic and disciplinary aspects of writing, suggesting an oral writing culture (Maagerø et al., 2021) and oral traditions of social studies teaching (Christensen et al., 2014;Lindh, 2019;Staf, 2019). This aligns with findings in my own research focusing on the teaching of grade 6 geography (Walldén, 2019) and history (Walldén, 2020b). The participant teachers emphasized oral interaction based on simplified texts that provided little opportunities for supporting the students' learning and expanding their linguistic repertoire. In the present study, I explore the context of the students' writing by putting their texts in relation to both the text material they studied and the literacy expectations conveyed by the written questions they answer.

Theoretical perspectives
In the present study, Systemic-Functional Linguistics (SFL) is employed to analyze the text material. SFL research has been instrumental in contributing knowledge about the linguistic tools necessary to express and to make meaning of subject-specific knowledge (Christie & Derewianka, 2010;Fang & Schleppegrell, 2010;Halliday & Martin, 1993). An important point of departure for SFL research is three metafunctions of language: the ideational function for representing human experience from different perspectives, the interpersonal function to create and maintain different roles and relationships, and the textual metafunction to manage the information flow in texts (Halliday & Mathiessen, 2014;Martin & Rose, 2007). The genre theory developed by Martin and colleagues (Martin, 1992;Martin & Rose, 2008) shows how linguistic resources for construing these meanings form register patterns that are coordinated by genres that fulfil different social goals.
In research on texts in history, an important finding is the learner's pathway (e.g., Coffin, 1997;Martin & Rose, 2008), which describes a progression of genres employing linguistic resources of chronology, causality, and rhetoric organization to different extents. Along this pathway, the representation of history is changed from concrete experiences (for example, recounts) to abstract explanations and arguments. Argumentative texts represent the most complex genre in the learner's pathway, requiring both developed field knowledge and rhetorical organization of arguments relating to the field (Martin & Rose, 2008). However, there is also a significant transition between chronologically unfolding recounts (through events or episodes) and genres foregrounding causal connections over temporal. The latter includes accounts that emphasize cause but still retain a temporal organization, and explanations that deal with multiple causes and consequences that are not temporally related (Martin & Rose, 2008, pp. 134-135). In the present study, the genre progression is used to contribute knowledge about the literacy expectations expressed in written questions the students were supposed to answer (RQ1). From the perspective of SFL, these genres become resources for engaging in disciplinary practices and express different aspects of historical thinking (see, Seixas & Morton, 2013), such as scrutinizing sources, discussing the historical significance of events and actors, acquiring historical perspectives, and reflecting on ethical dimension of history Furthermore, the genre progression is evident in the different register patterns they draw on (Marton & Rose, 2008, pp. 132-135). For example, accounts and explanations often use nominalized participants (e.g., Protestantism, the Reformation), while recounts tend to feature the "great men" of history (e.g., Gustav Vasa). In addition, accounts and explanations often use logical metaphors. This means that cause is often expressed within the clause, by processes (lead to), nouns (consequence), and prepositions (through) rather than by conjunctions (e.g., because, then). For young writers, there is also a progression between weakly specified coordinating conjunctions, (e.g., and, but-filling a range of possible functions) and conjunctions that express causal relations more explicitly (e.g., since, therefore, see, Schleppegrell, 2004, p. 67). Such explicit conjunctions are often unnecessary in temporally organized texts, which can rely on weakly specified conjunctions or implicit conjunction in sequences of sentences (Rose & Martin, 2012, pp. 278-279). In contrast, accounts and explanations require more explicit and specific conjunctions. Another linguistic feature associated with challenging texts in history is expanded noun groups, such as promises of forgiveness (see, Walldén, 2020b) and the degree of human ingenuity (Fang & Schleppegrell, 2010).
In relation to SFL, these linguistic features are resources for construing ideational meaning: experiential (e.g., participants, nominalizations) and logical (e.g., conjunctions, logical metaphors). Many of them correspond to general features of academic language perceived as important for the uncommon-sense writing required in formal schooling (Christie & Derewianka, 2010;Schleppegrell, 2004). They reflect the need to move along a cline between a spoken and written mode of communication (Halliday & Mathiessen, 2014). From the perspective of textual meaning, an important aspect is moving from using subjects and pronouns as themes to scaffold the unfolding of the text through marked themes (e.g., Christie & Derewianka, 2010;Martin, 1992, p. 440). Beyond the clause, the discourse semantic perspective of periodicity highlights information pulses in texts that guide the reader by introducing coming discourse and distilling accumulated new information (Martin & Rose, 2007). This use of compositional rhetoric-termed "power composition" in pedagogical metalanguage (Macnaught et al., 2013)-is important in texts relying on causality, such as the account and explanation genres discussed above. Examples of these information pulses are when the writer introduces an abstract phenomenon and then provides specific details, or when different factors in an explanation are summed up and put in relation to the abstract phenomenon introduced (Martin, 2013).
From the perspective of interpersonal meaning, recounts draw on evaluative resources of language to judge historical events or persons positively or negatively (Christie & Derewianka,(100)(101). They can also draw on narrative register patterns to engage the readers, which involves using evaluative resources of language to dramatize events and to underscore culturally valued moral positions (Christie & Derewianka, 2010, p. 40;Martin & Rose, 2008, p. 52). Accounts and explanations tend to be more dispassionate but draw on evaluative resources to appreciate the significance of historical events (Christie & Derewianka, 2010, p. 103). These ideational, textual, and interpersonal linguistic features will be used to analyze the students' written answers, both in relation to the literacy expectations of the written questions (RQ1) and properties of the text material studied (RQ2).
Beyond the linguistic analysis of text material, I share some sociological assumptions with other SFL-based research concerning how pedagogical practices can either perpetuate or remedy inequalities between different groups of students. Although some students can use linguistic tools to engage in disciplinary practices without explicit instruction, an equitable education must provide all students access to these features of language, including socioeconomically disadvantaged students and second-language learners that receive instruction in the majority language (e.g., Schleppegrell, 2013;Uccelli & Galloway, 2016). From the perspective of Bernstein (2000), who has been influential in the SFL research paradigm (e.g., Hasan, 1999;Rose & Martin, 2012), it is beneficial with a visible pedagogy that conveys literacy expectations in a clear way and supports the students in reaching them.
Research in linguistically diverse classrooms where many students are taught in their second language has shown the importance of an explicit attention to language, scaffolding sequences of teaching activities, careful selections of texts, and possibilities to engage substantially with content knowledge through various modes, such as writing, oral interaction, and visual representations (see, Hajer & Meestringa, 2020;Hammond & Gibbons, 2005;Nygård Larsson, 2018;Schleppegrell, 2013). This concern is shared in the present study, which focuses on writing practices in linguistically diverse classrooms located in a residentially segregated area.

Method and material
The present article is based on a classroom study of history teaching in Grade 6. One teacher and her two classrooms of Grade 6 students (40 in total) were followed for 12 weeks, spanning 32 lessons. My choice of school and participants was strategic (Thomas, 2011) because it aligned with my wish to research the opportunities created for students in disadvantaged areas to engage with and express disciplinary knowledge. I established contact with the teacher due to a shared interest in opportunities created for primary school students to engage with subject-related texts and discourse and was invited to her classroom to conduct research.
The school with the participant students and teachers is located in the disadvantaged part of a socially segregated area. According to official statistics, around 50 % of the students had a "foreign background", which means that they were either born outside of Sweden or had both parents born outside of Sweden. No specific information about the students' linguistic proficiency and length of stay in Sweden was collected. However, the teacher stressed that all the students, regardless of linguistic background, needed support in developing their ability to use resources of language and to interpret disciplinary texts in school. This was also expressed by several other staff members in the school. As the leader of a team of teachers, the participant teacher was a driving force in developing and employing a scaffolding pedagogy.
The studied teaching focused on the Vasa era in Swedish history, which belongs to the early modern period. It is considered a very important period in Swedish history. The scholarly literature on Gustav Vasa, the historical person mostly focused on in the teaching, stresses his role in centralizing power and building the Swedish nation state (e.g., Larsson, 2002). Relevant core content in the newly implemented curriculum, the wording is "The battle for royal power in the North. The Reformation and the emergence of a strong royal power in Sweden. Revolts and resistance against the royal power" (Skolverket, 2022).
A total of 37.5 hours of teaching was observed and documented by field notes and audio recordings (35 hours). 1 Two of the lessons were taught by a substitute teacher, who consented to participate in the study. An excerpt from one of those lessons is used to provide context for one student's writing.
In addition, 336 samples of students' writing were documented through photographs. Many of these were notes written about the students based on the content-related text material and images. This writing is illuminated in a previous publication (Walldén & Lindh, 2021). The analysis in the present article shifts the focus to written questions given by the teacher to the students and samples of students' written answers to five homework tests, given to them on a weekly basis, and a final test at the end of the content area. In the findings, samples of four students' writing will be particularly highlighted (focus students). To provide further qualitative insight into the conditions for writing, two of the texts used in the teaching and studied by the students were analyzed. 2 An overview of the data is presented in Table 1.
The process of selecting focus students for in-depth analysis was guided by a desire to capture the students' differing approaches and degrees of elaborations in the writing. This is further explained in the Analysis section. For this article, the written material was translated from Swedish to English by the author in a way which preserves the relevant linguistic categories (see , Table 2). For the sake of authenticity, the translations seek to preserve non-standard syntax and morphology. However, interpunctuation and spelling has been normalized in some of the texts since these presentational aspects may be unnecessarily distracting since they are not the focus on the analysis. Only the writing of the focus students was translated. In all the excerpts, wording in the original language, Swedish, is displayed underneath the English translations in small font.
From an ethical standpoint, informed written consent was collected from the teachers and the students' caregivers. The information included essential information, in accordance with the Swedish Council of Research (2017), such as the data collection procedure, expected publication forms, and the voluntariness of participation, including the right to discontinue participation. As written consent is problematic when caregivers use Swedish as a second or additional language, it was collected during a yearly student development discussion with the teacher. Thus, the caregivers were able to ask for and receive oral clarifications. The information letter itself had been adapted as far as possible to make it easier to read for Swedish language learners. Furthermore, I informed the students orally about the study and sought their continued consent to participate in the study (see, Tracy, 2010) by asking for permission each time I photographed a text. On occasion, the students showed reluctance to share their texts, which I, of course, respected.

Analytical procedure
Initially, the analysis was conducted according to the ethnographic aim of providing a rich understanding and description of the teaching studied (see, Fangen, 2005). Accordingly, I took field notes as a basis for more elaborated reflections on the writing practices (Duranti, 1997, p. 116), both in terms of the different literacy expectations carried by the written questions given to the students and of the students' written responses. These reflections were theoretically guided by the social semiotic lens employed in the study. In subsequent steps, the analysis was refined.
For the written questions, the categories presented in the findings (Table 3) were developed with inspiration from thematic content analysis (Braun & Clarke, 2006). The approach was appropriate since it allows interplay with different theoretical perspectives to develop relevant categorizations. All the questions were written in a table and coded according to the material in which they appeared, the sequential instance when the material was used, and the actual number of the question. 3 This was followed by an abductive categorization of the questions according to the literacy expectations they conveyed (Timmermans & Tavory, 2012). They were in part intuitive but also probed by theoretical connections, primarily to systemic-functional linguistics. The final themes were the question categories presented in the first part of the findings (see also, Table 3): These question categories and their distribution across the questions formulated in the text discussions, homework test and final test form the answer to RQ1. The practical procedure of the thematic analysis was to sort the questions in different table columns according to perceived differences in what they expected of the students. The themes developed in this process of familiarization (Braun & Clarke, 2006) were similar to the finalized categories but slightly more numerous. Notably, the finalized expand category first consisted of two: expand, describing relatively straightforward inferences, and explore, which went beyond the text. The distinction was not clear enough to be preserved in the finalized categories.
As previously mentioned, I developed the categories with inspiration from functional linguistic theory (Halliday & Mathiessen, 2014;Martin & Rose, 2007). For example, I made an analytical distinction between questions that merely prompted the students to provide short descriptions of historical persons (identify category) or events (reconstruct category, the latter pointing to linear recounts, and questions that required causal reasoning. The kind of question word (e.g., "what happened (reconstruct) . . . ", "why did . . . " (explain) was an important linguistic indicator. The expand category was evident through their use of mental processes (Why do you think . . .) to produce dialogically expansive questions, thus allowing many possible answers (Martin & White, 2005). From the perspective of reading comprehension, some of the categories also reflect the difference between literal and inferential meaning in texts (Rose & Martin, 2012), most notably between reconstruct and identify on the one hand (literal meaning) and expand on the other hand (inferential meaning).
The students' written answers to the questions were collected as part of the ethnographic fieldwork, in which I observed the writing. This meant that it was not practical, or even desirable, to collect all the samples of writing produced in the classroom. In total, samples were collected from 28 different students. However, I tended to focus on 13 particular students for a number of reasons. Firstly, these students were willing to give their continual consent; consequently, my photographing their texts became a part of the writing routine. It follows that I did not collect texts from students who showed reluctance or attended infrequently, for example, due to sickness. Secondly, I was interested in gathering texts from students whose texts differed with respect to, for example, degree of elaboration. Other points of variation included students using resources of learner language and multimodal representations. Next, the texts from the 13 students were analyzed in a more in-depth way. In the presentation of the findings, the number were further funneled down to four focus students showing different approaches to answering the questions. In accordance with RQ2, the analysis was conducted in light of the literacy expectations conveyed by the written questions the students answered. This means that the answers to some of the questions became more analytically interesting than others, such as questions that showed shifts in literacy expectations between the homework tests and the final test. In the linguistic analyses of the texts, I used the constructs described in the theory section. I particularly focused on how the students recontextualized the information in ways that reflected a written or spoken mode of communication. The transition between these modes is viewed as a cline in SFL research (Halliday & Mathiessen, 2014, pp. 726-728;Martin & Rose, 2007, pp. 298-300). Relevant features are displayed in Table 2.
Generally, the students did not go far down the written end of the cline. Answering several questions on a written test is conducive to a knowledge-telling model of writing (see, Bereiter & Scardamalia, 1978, pp. 9-10) in which strings of clauses representing content items are produced without much planning. However, the students' answers showed a variation with respect to the constructs listed above. The distinction between spoken and written mode must be viewed in light of the study's abductive approach. For example, a study based on samples of more developed writing, such as from older students, might have considered the use of subordination a feature of spoken language rather than of written language.
Finally, I selected two texts from the material the students studied (see Appendix A). The choice was directly based on the written questions and answers specifically highlighted in the other parts of the findings. In order to answer how the students' written answers can be understood in relation to properties in the text material they studied in preparation of the tests (RQ3), this analysis was also guided by the linguistic constructs in Table 2.

Findings
In the teaching, the students were assigned to study one part of the text material about the Vasa era every week. The texts were processed by question-based peer text discussions, the students' individual reading of the text while taking notes, and a written homework test that contained the same or similar questions as the ones used in the text discussions. However, the homework test was not given every week, due to the teacher's attendance and priorities. At the end of the content area, there was a final summative test. In preparation for this test, the teacher compiled and handed out all the questions used in the text discussions.
Two specific historical events will be highlighted in the findings. The first one, The Stockholm Bloodbath, occurred as part of Christian II of Denmark seeking to quell the resistance against Sweden remaining under his reign as part of the Kalmar Union by executing the followers of the fallen regent of Sweden, Sten Sture the Younger. It was a significant event leading up to Gustav Vasa's rebellion and eventual victory. The second event is the advent of Protestantism in Sweden, which played an important role in Gustav Vasa's successful endeavor to centralize power. Later sections of the findings will focus on how the above events are portrayed in both the text material studied by the students and in the students' written answers to questions relating to these events.

Shifting literacy expectations in the question material
This first section will take an overall perspective on literacy expectations in the content area conveyed by the written questions administered to the students (RQ1). Apart from the events described above, these questions pertained to events before the Stockholm Bloodbath, Gustav Vasa's war with Christian II, and various uprisings against Gustav Vasa.
The analytical process resulted in the seven categories outlined in Table 3. Figure 1 shows the distribution of question types in the different materials used in the teaching: questions for peer group text talks, homework tests, and the final examining test.
As indicated by the table, the text talk questions were numerous (29) compared to the homework test (19 questions) and to the final test (9 questions).
The identify question type requested descriptions of historical characters. The type occurred just twice in both text discussions and homework tests, and once in the final test ("Who was Sten Sture"). The category was similar to the more common reconstruct question type. It requested the retrieval of specific information from the text that related to historical events. This was the most common question type in the text discussions (9 of 29) and in the homework test (5 of 19). It was also represented in the final test (3 of 9). From the point of view of literacy expectations, both these questions merely require using information stated in the material. In other words, they did not necessitate inferences, interpretations, or understanding, or causal logical connections. The rephrase question type mainly required explanations of the meaning subject-related vocabulary, such as hereditary monarchy, dynasty, and invade. In answering these questions, the students had to rephrase these words in common language (for example, "What does it mean that Gustav Vasa introduced hereditary monarchy?"). Two examples from this category were different, as they involved rephrasing something stated in common language by using information in the text ("What does it mean that the party finished in a bad way?", which is explained in the next paragraph of the text) and by explaining an expression in the textbook material relating to the trustworthiness of historical texts ("What's the difference between history and stories?"). 4 The latter example occurred in a text discussion and was the only question in the material pertaining to evaluating a historical text. Aside from the "history and stories" example, the rephrasing questions pointed to low level literacy expectations since they concerned understanding single words and expressions with meanings quite easily retrieved from the text.
In contrast to the reconstruct category, the expand questions required the students to infer meaning not explicitly stated in the text. One such example followed a reconstruct question about Christian II's invasion of Sweden: "How do you think this happened?" The invasion was not explicitly described in the text, which prompted the students to make use of related knowledge. Similarly, the students were prompted to infer characteristics of Gustav Vasa by commenting on a famous image of him. One of the expansive questions related to the Stockholm Bloodbath, "Do you really think that Sten Sture's men had wrong ideas of Christendom (heresy)", was leading in nature, while a question related to the monarch's family tree asked the students why they thought Gustav Vasa had three wives. Both these examples prompted the students to think about the motives of historical characters. The question type was mainly represented in the text discussions and in the homework tests. Indeed, the teacher expressed that she sought to stimulate the text discussions by these dialogically expansive questions (see, Martin & White, 2005). One of these was included in the final text: "Why do you think there was such a big difference between what Gustav Vasa got to learn when he was young compared to what you get to learn?" More explorative in nature, it required the students to connect information about the historical actor and period to their own experience.
The fifth and final major category, explain, required students' causal reasoning about historical events. A question posed in a text discussion, "If Christian II had continued as king over Sweden, how do you think it would have affected us today?", drew attention to long term consequences for the nation. Another text discussion question, "Why did farmers in Småland revolt?", required the students to come up with relevant information closely connected to the event referenced. A final test question, "How were things in Sweden after the Reformation?", could refer either to shortterm or long-term effects. In two final test questions, the teacher used several metalinguistic expressions to accentuate the expectation of causal reasoning, for example, "Tell us about the Stockholm Bloodbath (think: Why? Who? How? Consequences?)". 5 Another final test question was similarly constructed: "Tell us about the Bell Rebellion (think: why, who, how)". Although rare in the text discussions (3 instances) and homework tests (1 instance), the explain category dominated the final text (5 out of 9 questions). This shows that the literacy expectations were scaled up in the final test, with less emphasis on identifying and reconstructing and more emphasis on explaining.

Teaching through stories: the textbook material
Before we turn to examples of how the students negotiated the literacy expectations in their writing, we will consider two texts that the students had discussed and studied before answering questions in the homework test and in the final test. The texts are from two different teaching resources and have been illuminated in a previous publication (Walldén, 2020a). The first text is about the Stockholm Bloodbath (Moberg, 2019, see, Appendix A). It was selected because it relates to a test question shifting from reconstruct to explain type in the final test.
The first part foregrounds a person, Gustav Trolle, who instigates the Stockholm Bloodbath. This is evident through several unmarked themes constituted by the name or a pronoun referring to the person and a focus on this historical person's reasons and actions: "Gustav Trolle was . . . He had . . . ., Gustav Trolle was . . ., Now, he had . . . ". The event is named in the heading to the second part of the text: "The Stockholm Bloodbath". This part foregrounds aspects of time and place by several marked themes: "On November 8 . . . In the Grand Square . . . For three days . . . First in line . . . ". A concessive conjunction ("Even . . . ") is used as an evaluative resource to mark the grisly detail of Sten Sture's corpse being dug up and burnt, while the final clause, "And blood flowed on the streets of Stockholm", refers to the heading naming the event and unpacks its meaning by representing the nominalized event (bloodbath) as a process ("blood flowed"). This further accentuates the drama. Except for the heading, the name of the event is not used as an abstract participant in the text.
A following part of the text material states that people revolted against the king, Christian II, and that they were led by Gustav Vasa, whose father had counted among the victims of the bloodbath. However, this is not explicitly stated as a consequence of Stockholm's Bloodbath. Furthermore, the connection between the event as instigated by Gustav Trolle and revolt against the King is not explained. For these reasons, the text gives very little support for understanding or explaining the event in terms of causes and consequences. It follows that the text does not stress the historical significance of the event. Instead, the text describes the event and persons involved in the Stockholm Bloodbath, while conveying some of the drama.
The text about Sweden becoming Protestant (Grauers, 2000, pp. 10-11, see, Appendix A) is longer and provides more explanation. This is indicated by several marked themes and conjunctions that explain factors fueling the reformation, such as Gustav Vasa's need for money for his war efforts (e.g., "therefore had to", "Furthermore . . . "), how the general poor state of Sweden contrasted with the wealth of the Church (" While . . ., "In the churches . . . ", "Moreover . . . "), and Gustav Vasa's inability to access the riches of the Church (" . . . in order to pay Lübeck . But he did not get to"). In addition, the text explains how Protestantism enabled him to gain power: "The Protestants did not want the Pope as the leader. Instead, the King of the country would decide . . . ". However, the text follows a narrative rather than explanatory structure (Martin & Rose, 2008) by providing an orientation (describing the poor country at the time), a complication (Gustav Vasa's immediate need for money, obstructed by the Pope), an evaluation (the King's emotional reaction to the event, "got furious"), and the final solution (Protestantism). The phrasing of the solution is a clear example of the text using spoken language syntax and vocabulary, as meaning is spread loosely across a clause complex: "Gustav Vasa thought that sounded good and because of that he wanted Sweden to come protestant". The name Gustav Vasa or pronouns referring to him-one of the "great men" in Swedish history-occur frequently as a specific participant, which shows that the texts focused on his doings rather than on the process of Sweden becoming Protestant.
In the following paragraphs, concrete changes due to the Reformation are described, such as "They [people] did not find the churches as beautiful without the wall paintings and silver treasures were sent to the king to be melted down" and "the preaching got much longer". Thus, the consequences were described in terms of observable changes, while the long-term significance, such as the centralization of power in Sweden, was not made clear.
Although the text relating to Protestantism is more elaborated than the one about the Stockholm Bloodbath, the two texts share features of recounts and narratives rather than of causally organized texts, such as accounts and explanations. Although drawing on story patterns to make content-related texts available to young learners is common and not necessarily problematic, there is a tension between these texts and the scaled-up literacy expectations of causal reasoning in the final test of the studied teaching. The coming sections will focus on how the students negotiated these expectations.

Students' negotiating the shifting literacy expectations
In this section, I focus on how the students negotiated the scaled-up literacy expectations. My point departure will be the increased focus on explanatory writing in the final test.

Shifting from reconstructing to explaining
The questions about describing the Stockholm Bloodbath are particularly interesting since it occurred as a reconstruct question in the homework text and was scaled up to a explain question in the final test. The excerpt below shows how a student answered both these questions. Small font shows wording in original language.

Excerpt 1, Student A's answer about the Stockholm Bloodbath
Student A wrote answers on separate lined papers instead of the question sheet, showing a commitment to this presentational writing. The answer to the reconstruct question starts by stating the location of the event and progresses by describing chronological events: Sten Sture's men being accused of heresy followed by their beheading and the blood flowing "all over the grand square". In describing the beheading, the student uses a dependent clause with a conjunction, "Since they were beheaded". Although this appears redundant, it nevertheless indicates an ambition to use causal reasoning. In rounding off the answer, the student re-states the event, "and that's the Stockholm Bloodbath", which signals the completion of the answer by using the event as an abstract participant. Both the use of a conjunction and the summing up of the statement indicate the student striving to adapt rhetorically to a written mode of communication.
The answer to the final test explain question is different. However, in contrast to some of the other students' answers (see below), it is shorter rather than longer (47 words compared to 53). The student begins the answer by describing the accusation of heresy. Unlike the answer to the reconstruct question, it is followed by an explanation of both the meaning of heresy and the consequences of being accused of heresy: "and if you have done heresy that is wrong thoughts about Christianity you get beheaded". The use of a causal ("if") and a reworking ("that is") conjunction shows how the student adheres to the literacy expectation of explaining. Unlike the reconstruct answer, the student mentions Gustav Vasa's family being among the victims. To conclude the answer, the student uses one of the parenthetical metalinguistic prompts in the question, the logical metaphor "consequence", as an abstract participant to describe a related result of the bloodbath: "The consequences to this [DM]  repercussions for Swedish history. Thus, the answer is rhetorically organized in a way that reflects a written mode of communication and aligns with the literacy expectation of explaining.
In summary, the answers show the student navigating the scaled-up literacy expectations with accuracy and relative success. In both answers, the student uses a written mode of communication by managing the information flow and being explicit with conjunctions, while adapting this mode to the different purposes of describing and explaining. This is noticeable in the way the student initiates and concludes the answers. This also shows that the student took different perspectives on the field in a way which corresponds to the shifting literacy expectations. This was achieved by including more information about locales and events in the reconstruct answer, while stressing information about the heresy and Gustav Vasa's family to support the causal reasoning necessary to provide an answer to the explain question, especially the "why" and the "consequences".
The example below shows a student employing a different approach to the same questions.

Excerpt 2, Student B:s answers about the Stockholm Bloodbath
In the reconstruct answer, the student starts with briefly stating the context of the event: people being executed because of heresy. Then, the student accounts for the order in which people from different estates were executed. Finally, the student presents the gruesome detail of the digging up and burning of Sten Sture's corpse. The detail was accentuated in the text material and picked up by several of the students in their answers. However, Student B did not describe the political context of the event, including the reason for accusing people in Sten Sture's party of heresy or information about who delivered the accusation. The use of "exactly" in the phrasing "exactly due to heresy" likely follows the construction of the homework test: the preceding question asked the students to explain the meaning of heresy. In contrast to the text material, the student uses active voice and a pronoun ("they") to describe the burning of the corpse, which reveals a lack of information about who committed the act (also not stated in the text material). Overall, the lack of both elaboration and explicitness reflects a spoken mode of communication.

Explain question in final test
The answer to the explanatory final test question is substantially longer (98 words compared to 38). The student develops the field by adding information about the people involved and the party at which Gustav Trolle delivered his accusation. In contrast to the reconstruct answer, the concept of heresy is explained in parentheses. The student concludes the answer by stating what happened during the bloodbath itself, in a similar phrasing as in the reconstruct answer. Furthermore, the student adds to parenthetical comments of Sten Sture's men being at the party and, once again, the burning of Sten Sture's corpse. This indicates an ambition to substantially develop the answer compared to the homework test.
The use of information and wording from the text material enables the student to provide some explanation of the events leading up to the bloodbath, for example, by mentioning the enmity between the fallen regent Sten Sture and Gustav Trolle. However, the student uses a chronological recounting structure spreading meaning throughout several clauses with repeated use of "he" as an unmarked theme: "He was one of Sten Sture's greatest enemies. He had not forgotten that he was cast into prison . . . He closed the door". Similar to the text studied, the student expresses causal relations by implicit conjunctions and vague coordinating conjunctions (and, but), rather than conjunctions used in causal reasoning (e.g., since, because). This spoken mode of writing, further indicated by a lack of interpunctuation and addition of information in parentheses, gives the reader little guidance in retrieving important information from the text. The detail about Sten Sture's corpse, which concludes both answers, mirrors the teaching material reliance on story patterns to account for events and for adding drama. It gives no information about important consequences of the event.
Among the students, the relative success of Student A in fulfilling the literacy expectations in the final test can be attributed to a more independent stance to a text material previously shown to be focused on describing and dramatizing the persons and events rather than on explaining the historical significance of the Stockholm Bloodbath. While Student B's answer to the explain type questions builds closely on the text material and retains many of its linguistic features-such as unmarked themes, specific participants, and reliance on temporal rather than causal organization -Student A chose a rhetorical organization that highlighted an important consequence of the event. Furthermore, Student A adjusted the rhetorical organization to the different demands of reconstructing and explaining the past. In contrast, Student B responded to the scaled-up literacy expectations by reconstructing more information about persons and events from the text material. This was the case for many of the students.

Explaining reasons for the Reformation
Next, we turn to how the students answered explain type questions relating to the Reformation of Sweden. These did not occur in the homework tests 7 but were posed to the students in the final test. As discussed in the previous section, the text material provided concrete reasons for Gustav Vasa's promoting the Reformation (for example, needing money to pay a debt) and mostly described straightforward changes after the reformation, such as people missing "nice wall paintings" and having to endure longer church services. The analysis of the students' texts shows different approaches and strategies to answer the explain type questions. Below, we show four different answers to the following explain type question: "Why did Gustav Vasa wish Sweden to become Protestant?"

Excerpt 3, Students answer about Protestantism in Sweden
Although all the answers mention the debt to Lübeck, they use different textual and logical organization. Both Student A and D start with Gustav Vasa's wish to rule over the Church, followed by the reference to the debt to Lübeck. Student A mostly uses coordinating conjunctions ("and . . . ") but also two logical metaphors ("due to" . . . "through") to provide the explanation. Once again, the rhetorical organization of the short answer appears planned since it both begins and ends with explaining the role of Protestantism during Gustav Vasa's reign. This is achieved through using Protestantism as an abstract participant ("wanted Sweden to become Protestant because . . . ") and a circumstantial element supported by a logical metaphor ("through Protestantism he could . . . ").
Student D initiates his answer similarly-he begins with stating Gustav Vasa's wish to rule the Church. In providing the explanation, he uses two congruent causal conjunctions, "because then" and "so". However, unlike Student A, Student D does not explain how Protestantism enabled Gustav Vasa to rule over the Church and take money. Furthermore, there is no mention of the political roles involved: how the King's power relates to the Pope's. Thus, the field is not as developed in terms of processes and participants. These features reflect a difference in mode, with Student A's more developed and rhetorically organized answer indicating a written mode of communication to a higher degree than Student D's. Despite evident second-language features, such as lack of subject in some clauses (then [he] could rule . . .), the answer clearly adheres to the literacy expectation of explaining by selecting relevant information and successfully using conjunctions. Both Student A and Student D move from a general concern (the monarch accumulating power) to a more specific and immediate one (paying a debt to finance a war).
Students B and C make a different textual choice by foregrounding information about the debt to Lübeck. In doing so, they adhere more closely to the text in the teaching material. While both answers start by stating the need for money, they do it in different ways, which signal a difference in mode. Student C's wording is much more precise (" . . . because he and the kingdom needed Student A GV wanted Sweden to become protestant because then the King ruled over the Church and not the pope and GV needed money due to his debt to Germany and through Protestantism he could take money from the Church.

Student D
He wanted to makes Sweden Protestant because then [he] could rule over the churches in the country and take money from the church so [he] can pay back to Lübeck in Germany.

Student C
He wanted that because he and the kingdom needed money to pay back to Lübeck. Because of having "rented" soldiers and war ships when he surrounded Stockholm. And within Protestantism. the King is the one with most power in the church instead of the pope. And if he rules, he can [DM] just take valuable treasures. And melt down. And send to Germany.

Student B
He wanted that just because of money. He had heard about Protestantism. He had heard that there were lots of riches in churches & monasteries. And he had debt to Germany because they help him win Sweden. Despite all he taxes he got it wasn't enough. He could not get to take the riches because the highest boss in monasteries and churches was the pope in Rome. He did not let Gustav V take them.
Han ville bara det för pengar. Han hade hört om protestantismen. Han hade hört att det var många rikedomar i kyrkor & kloster. Och han hade skuld till Tyskland för dem hjälp han vinna Sverige. Trots all skatt han fick räckte det inte. Han fick inte ta rikedomarna för den högste chefen i kloster och kyrkor var påven i rom han lät inte Gustav V ta dem. money") than Student B's ("just because of money"). Student C shows several further signs of adapting to a written mode of communication by giving relevant background information to the debt ("having 'rented' soldiers", "surrounded Stockholm") and accounting for the King's position in Protestant churches through an expanded noun group ("the one with most power"). While the student relies on spoken language coordination (and . . .), rather than on more explicit causal conjunctions, the information flow is facilitated by a marked theme containing a key term ("within Protestantism") and a fronted subordinate clause ("if he rules . . . "). Student B's answer appears less planned, largely due to repeated unmarked themes ("he just wanted it . . . He had heard . . . he had debt") and sparing use of explicit conjunctions. The student uses the conjunctions "because" and "despite" to give some background information but does not integrate this information with the role of Protestantism. For example, the relevance of "He had heard about Protestantism", a textually foregrounded statement because it comes early in the answer, is not explained in the following text. Furthermore, the student uses resources of everyday language to convey that "the highest boss in monasteries and the churches was the Pope in Rome" and that the Pope "did not let [Gustav Vasa] take [the treasures]"; however, the student does not explain how this power would be changed by Protestantism. Therefore, the key power relations at play need to be inferred from the answer. This reflects a spoken mode of communication, one that is further underscored by a lack of interpunctuation. Although the answers of Student B and C are quite different, they both focus on Gustav Vasa's debt to Germany, which is also emphasized in the text material. As such, these students, with differing success, seem to use a reconstruct approach in their explanatory answers. This was not the case with Student A and D, who gave shorter answers foregrounding the issue of power.
By adhering to a written mode of language use, Student A and C appear to be most successful in their explanatory answers, although in different ways. Student A formulated a succinct response by selectively drawing on the text material and using several conjunctions, while Student C achieved the explanation by incorporating information in the text material in a planned way. Student C's reliance on the textbook material appears less successful due to the unplanned nature of the response. Despite learner language-related hurdles, Student D's answer shows a tighter integration of relevant meanings through textual choices and use of conjunctions.

Explaining effects of the reformation
In the final part, we turn to how the students answered a second explain-type question regarding the reformation: "How were things in Sweden after the Reformation?" The answers to this question showed comparatively little variety since they mainly repeated the concrete, observable changes stated in the simple text material. An illustrative example from a (non-focus) student is shown below: Excerpt 4: A student describes changes due to the Reformation of Sweden: All the wall paintings and all the silver disappeared, benches they had got. The preaching was in Swedish and much longer. And they talked a lot about behaving in church.
The answers differed somewhat in the information included. For example, Student A referred to the difference in atoning one's sins: "you don't need to pay for your sins but just regret them for real". Like several other students, Student B wrote a long answer, which included Martin Luther's thoughts about Christianity and his role in the reformation, thus incorporating information in the text material not particularly relevant to the explanatory question at hand. This further shows the difficulty in employing reconstruct approaches to the scaled-up literacy expectations. Student C selected similar information, as did other students, to describe changes, but chose a multimodal representation instead (approximated in Figure 2) The visual organizer efficiently represents the concrete changes described in the text material and reproduced more conventionally in verbal text by the other students. An elaborated explanation of long-term effects would have required a different representation, but such an answer is not prompted by the text material.
Another outlier response to the same question was produced by Student D. While all the other students adhered closely to the text material, Student D used meaning conveyed orally by a substitute teacher responsible for the lessons about the Reformation. This teacher brought a different book to the classroom and elaborated through a short read-aloud activity that took place after the classroom reading of the primary text material. A short excerpt from the activity is shown below:

Excerpt 5: Substitute teacher's oral elaboration on the text material
What benefit did people have of monks and nuns? Well, they were of great benefit./. . ./That is, monks and nuns were educated, weren't they? You see, they could read, and not everyone could do that in those days. So, when they closed the monasteries and they moved out. Then, those who had taught children to read disappeared. Then, those who were good at taking care of sick people disappeared. And monks and nuns, they could cultivate herbs that could be used to produce medicine. So, these things got much worse for the country.
This oral re-telling expanded significantly on the text and is reflected in Student D's answer to the test question about the reformation. The answer shows L2-related features of syntax, notably some missing subjects and finites. Furthermore, the student struggles with combining information from the text material ("found two new priests") with the information about monks, nuns and monasteries conveyed orally by the substitute teacher. Together with unclear evaluative phrasings introducing and concluding the answer ("In Sweden [people] think it was better" . . . "Which [for] Sweden [was] not good"), this results in a lack of coherence. However, in the middle part of the answer, the student successfully conveys how the closing of the monasteries impacted the country. The student uses a marked temporal theme ("After the reformation . . . "), a fronted temporal subordinate clause ("when [they] took down . . . "), and a causal conjunction together with resources of appreciation ("It was bad because . . . ") to guide the reader through the explanation. Thus, the student partly succeeds in using the teacher's oral elaboration in answering the explain type question. The selected outcome arguably carries more significance than some of the changes described in the text material and reproduced in other students' answers, such as the insertion of church benches and the removal of church wall paintings. Unlike the answer shown in Excerpt 4, the student uses evaluative resources of appreciation to evaluate the event, which is a sign of adjusting to the scaled-up literacy expectations.
That none of the students' answers elaborate on long-term motivations and effects of the reformation in their explanatory answers is not surprising, since these more abstract and linguistically challenging perspectives have not been offered in the teaching. However, the analysis has shown different approaches by the students to navigate the scaled-up literacy expectations. These include selecting relevant content from the text material and using conjunctions to produce relevant explanatory meaning, maximizing the possibility to reproduce meaning in the text material (to varying degrees of success), employing visual modes of representation, and making use of disciplinary meanings conveyed orally. Several of the more successful strategies involved students distancing themselves from the simplified text material.

Discussion
The findings have highlighted how students negotiate shifting literacy expectations in the teaching of history. The research has been motivated by the need to explore and critically discuss the possibilities created for students in socioeconomically disadvantaged areas to engage with and express content knowledge through reading and writing.
In answering RQ1, the present study contributes to previous research on writing in history teaching by showing how different literacy expectations were expressed in written questions administered to the students in a final test and in activities leading up to the test. The categories developed in the findings show that the students were not expected to engage in the kind of disciplinary sourcing and argumentation emphasized in current research (Fitzgerald, 2019;Rantala & Khawaja, 2018;Wissinger et al., 2018). Except for a question used in a text discussion, the students were not required to consider texts as sources (cf., Shanahan & Shanahan, 2008;Wineburg, 1991). Furthermore, there were no expectations of the students to produce written arguments-the genre considered most complex along the learner's pathway (Martin & Rose, 2008). In accordance with Bharath and Bertran (2015) and Olvegård's (2014) textbook studies, argumentative texts were not featured in the text material used in the teaching.
Nonetheless, the comparison of questions administered to the students showed that the literacy expectations were scaled up in the final test since most of the questions belonged to the explain category, which required the students to reason in terms of cause and effect. The homework tests and text discussions instead emphasized the reconstruct category, which invited temporally organized recounts. Thus, the final test required a step forward along the learner's pathway. This adds to the view that expectations of disciplinary writing in the Swedish national curriculum might be more evident in summative assessment situations than in on-going teaching practice (Molin & Grubbström, 2013;Samuelsson & Wendell, 2016;Staf, 2019). While previous research has shown how history textbook material relates to the learner's pathway (Bharath & Bertran, 2015;Olvegård, 2014), the present study has added a new perspective by highlighting how literacy expectations can shift within the teaching of a specific content area.
The categories developed and described in answer RQ1 also show potential disciplinary writing practices aligning with important aspects of historical thinking that could be promoted in the teaching (see, Seixas & Morton, 2013). The judge category was represented merely by one question in a text discussion. It can be a fruitful point of departure for writing that explores ethical dimensions of history through considering the actions of historical actors in light of present-day concerns. Furthermore, questions of the expand type can promote students' taking of historical perspectives by thinking about the motives of historical actors and connecting the lives of peoples of the past and their own lives and experiences.
RQ2 and RQ3 are closely connected since they put the students' texts in relation to important contextual aspects of the writing: the previously mentioned literacy expectations conveyed by the written questions (RQ2) and the text material the students studied in preparation for the tests (RQ3). Although all the students focused on in the study produced writing characterized by features of spoken language, such as frequent use of implicit or weakly specified conjunctions, they showed different approaches to negotiating the literacy expectations. To varying degrees of success, Student B and Student C relied on temporally rather than causally and rhetorically organized writing to answer questions in the explain category. This approach seemed directly prompted by both the temporally organized text material studied and the emphasis on reconstructing historical events in the text discussions and in the homework tests. In contrast, Student A more clearly adjusted to explanatory writing by using features highlighted as important for disciplinary writing in previous SFL-based research. This includes logical metaphors and using rhetorical organization, "power composition" (Macnaught et al., 2013), to emphasize consequences and abstract concepts. This kind of writing was neither modelled in the teaching nor promoted by the text material.
It is not surprising that other students relied on the reconstruct approach and expressed perspectives on the content similar to the text materials. However, there were important differences in how developed their answers were in terms of logical connections and use of participants. Student B's reliance on features of spoken language-such as repeated unmarked themes, implicit or weakly specified conjunctions, and lack of generic, content-related participants-stands in sharp contrast to the literacy expectations of secondary education that the student will soon enter. Student D's writing illustrates the dual challenge involved in expressing content knowledge in a second or additional language (Schleppegrell, 2013;Uccelli & Galloway, 2016). The student shows signs of the rhetorical and logical organization of text conducive to explanatory writing while also struggling with basic syntactical structures in the target language, such as mandatory finites and subjects in clauses.
The findings show the importance of all students being sufficiently scaffolded to develop the linguistic resources necessary for participating in disciplinary practices (Rose & Martin, 2012). Although the teacher made several efforts to support the students, for example, by promoting text discussions, the increased literacy expectations first evident in the final test is a sign of an invisible pedagogy (Bernstein, 2000). Therefore, the present study adds to concerns expressed in earlier Nordic research about the lack of attention to metalanguage and knowledge-developing aspects of writing (Christensen et al., 2014;Lindh, 2019;Maagerø et al., 2021). If the students participating in the current study had received the opportunity to engage in writing as a collective endeavor and participate in scaffolding exchanges that involved explicit attention to linguistic features used in explanatory writing, such as in the joint deconstruction and construction of texts (Fenwick & Herrington, 2021), the differences between them would likely have been less striking. Previous research has shown the significance of students' receiving instruction that develops strategies for structuring their texts in alignment with disciplinary goals (De La Paz et al., 2017;Finlayson & McCrudden, 2020). However, this also requires the teacher to choose texts that are appropriate models for the genres and registered targeted in the teaching, which, in turn, demands an awareness of disciplinary texts and language that many teachers are lacking.
In understanding disciplinary writing practices, it is important to consider the resources available to the students beyond the verbal texts (Hajer & Meestringa, 2020;Hammond & Gibbons, 2005). For the students, one literacy expectation seemed very clear: that they were supposed to learn and reproduce content items from the unchallenging text material studied. Indeed, the progression through the content area was, to a large degree, the same as the weekly progression through the text material by notetaking, text discussions, and homework tests. However, in answering the final test question about consequences of the Reformation, two students chose to build their answers on other modes. Student C elegantly captured the simple changes described in the text through a visual representation. This indicates that Student C linked the test-related writing with prewriting activities in which multimodal representations were encouraged (see, Walldén & Lindh, 2021). Student D used a teacher's oral elaboration on a text that was not included in the printed text material the students studied. Although it proved challenging for the student to transduce the teacher's elaboration in writing and connect it to information in the text material, the perspective on the question was apt.
These latter approaches show how reading and writing can interplay with listening and use of visual resources in a way which expands the possibilities of meaning making in disciplinary literacy practices (e.g., Bezemer & Kress, 2016;Ivanić, 2004). However, most students seemed to follow a narrow path of least resistance in their test-related writing that was charted by the meagre text material. In relation to previous research (Rantala & Khawaja, 2018;Wineburg, 1991), this gives further evidence of students' tendencies to trust textbooks to give then trustworthy and exhaustive information about the matter studied. Particularly in the teaching of history, such perceptions should be challenged rather than perpetuated. With the "factual" turn evident in the recently implemented Swedish curriculum, with less emphasis on critical and interpretative practices particularly in elementary grades, this seems less likely to happen. In light of the present findings, the removal of historical and metalinguistic concepts as part of the core content-such as the meaning of chronology, cause, and consequence-appears equally problematic because such concepts are important for the enactment of a visible and scaffolding pedagogy.

Conclusion
The present study has shown the significance of teachers taking inventory of literacy expectations in different activities, both regarding the alignment between these activities and their connection to specific aims for learning in history. The study also reveals that the students' writing is interconnected with their possibilities to engage substantially with content language through reading and various other modes. It is important to note that the scaffolding and language-conscious pedagogy proposed in this article, in alignment with other SFL-based research (e.g., Fenwick & Herrington, 2021;Llinares & McCabe, 2020;Moore et al., 2018), would likely benefit all students, not just the ones most evidently in need of support. For example, Student A tackled the literacy expectations with relative success but should be challenged to consciously develop the emerging rhetorical structures in relation to more substantial and challenging text material. Thus, the student could gain from the same metalinguistic scaffolding, based on high-quality texts, given to students to develop mastery of the target language alongside the capability of expressing specific content knowledge.

Limitations of the study and suggestions for future research
The present qualitative study is limited to investigating literacy expectations and students' writing in two grade 6 history classrooms taught by the same teacher. Thus, the findings relating to categories of literacy expectations and students' approaches to writing are not statistically generalizable. Also from a qualitative standpoint, a larger study involving more teachers and students would have enabled more robust categories and illuminating comparisons between classrooms. However, the study has been able to provide rich exemplary knowledge (Thomas, 2011) about students' navigating shifting literacy expectations in a specific content area. In future research, the study's novel approach of charting literacy expectations and analyzing students' disciplinary writing in light of both test-related literacy expectations and text material can be used to explore disciplinary literacy practices across content areas and stages of schooling. In particular, it could be applied to less-researched content areas of social studies teaching, such as civics and religion.