Internal consistency in a Swedish history curriculum: a study of vertical knowledge discourses in aims, content and level descriptors

ABSTRACT The study presented in this article examines the internal consistency of the formal Swedish history curriculum for the upper-secondary school. The research question addresses the extent to which disciplinary aspects—also referred to as vertical knowledge discourses—are transferred from fields of knowledge production to the Swedish formal history curriculum. A qualitative research design is adopted and document content analysis is used to examine the presence of vertical knowledge discourses in the aims, the core content and the level descriptors of the curriculum. The results show that a vertical knowledge discourse is present in two areas of the curriculum: the aims and the section that defines the content of history education. An exception is identified in the section of the curriculum that focuses on assessment, the knowledge requirements. Here, the presence of vertical knowledge is marginal. The conclusion is drawn that there is an inconsistency regarding the presence of vertical knowledge discourses in the Swedish history curriculum. This inconsistency can diminish the availability of vertical knowledge for the students and also lead to a situation where control over history education to some degree is transferred from curriculum designers and teachers to other agents.


Introduction
Assessment of student knowledge is an important feature in many educational contexts. Any assessment of student knowledge always relies on a set of levels related to the students' knowledge, and such levels can be more or less explicitly formulated. In 1994, Sweden introduced explicit level descriptors in the curriculum as a part of a shift from a norm-based to a standards-based grading system. This shift has been described as a part of an international trend towards standardsbased curricula (Sundberg & Wahlström, 2012). Introduced in the mid-1990s, the Swedish standards included explicit descriptions for three grading levels, a feature that is also present in the curriculum that has been in effect since 2011. These explicit descriptors for three levels make the Swedish curriculum more detailed in this respect than curricula in most other countries (Lundahl, Hultén, & Tveit, 2016). Another characteristic of the Swedish curriculum is that the subjects are based on subject-specific competencies, while most other standards-based curricula are based on more generic competencies (Sundberg & Wahlström, 2012). The combination of a subject-specific approach and explicit level descriptors is interesting. It is claimed that one important feature of a high-quality curriculum is that its components are consistent with each other (Akker, 2006;Ercikan & Seixas, 2011). From a consistency perspective, it is important to examine the Swedish curriculum because it combines a subject-specific approach with explicit level descriptors. Many models for addressing descriptions of levels are generic (Anderson, Krathwohl, & Bloom, 2001;Biggs, 1982;Bloom, 1956), and as their use in relation to subject-specific content is not unproblematic (Wineburg & Schneider, 2010), the aim is to examine how the formulation of explicit level descriptors is addressed and how they are related to the subject-specific context of the curriculum.
I will approach the issue of consistency in the Swedish formal curriculum by examining the formulations in one specific subject-history-and the relationship between three sections of the Swedish history curriculum. Two of these sections are directly related to teaching: the first includes the aims of history education, whereas the second defines the core content that history education should address. The third part of the curriculum that is of interest for this study contains the level descriptors and is therefore more related to assessment. In this third section, the descriptors are listed for each of the three levels that the teachers are required to use when evaluating and grading their students' knowledge.

Theoretical considerations
To examine how subject-specific competencies are combined with explicit level descriptors in the Swedish history curriculum, I will use concepts from both curriculum theory and history education research. From curriculum theory, I draw upon a number of concepts formulated by Basil Bernstein: fields of knowledge production, classification, framing and verticality. From history education research, the discussion on firstand second-order concepts is the most relevant.
The subjects in which teachers assess students' knowledge are in many cases related to a scientific discipline. Within these disciplines, specified knowledge and ways of thinking have been developed over time. In the historical discipline, this can be exemplified with how evidence is constructed by interpretations of historical sources (Berkhofer, 2008;Jordanova, 2006), and such specified ways of relating to the world cannot normally be acquired outside a society's learning institutions. Bernstein (2000) refers to this kind of specialized knowledge as a vertical knowledge discourse. The everyday kind of knowledge that individuals can acquire outside school is referred to as a horizontal knowledge discourse. The advantage of a vertical knowledge discourse is that it provides individuals with knowledge that makes it possible to perceive and engage with the world in ways that would be impossible without it. The knowledge and ways of thinking that are developed within the scientific disciplines are transformed in some ways when being transferred into school curricula. This transformation process is labelled as a recontextualization of the disciplinary knowledge (Bernstein, 2000).
In a discussion on the principles of curriculum content, in the wider meaning of the word, Young argues that vertical knowledge in a curriculum must be rooted in scientific disciplines and address the epistemic grounds of these disciplines (Young, 2013b, pp. 196-197). The argument for connecting curricular knowledge to the epistemological structures of the disciplines is that they help students not only obtain access to knowledge that is generalizable but also use such knowledge to 'think the previously un-thought' and thus understand the world in ways that would not have been possible without formal education (Young, 2013a, p. 246).
Muller argues that there are two dimensions of each vertical discourse: the products of an academic discipline and the procedures necessary for the construction of the disciplinary product. He presents two kinds of knowledge within the latter dimension: inferential knowledge and procedural knowledge. The former, inferential knowledge, consists of the processes that are applied when elements of a discipline are assembled to more coherent structures. The latter, procedural knowledge, consists of the procedures that are used in order to establish empirical statements (Muller, 2014, pp. 261-263). If the content in a curriculum is meant to contain vertical knowledge structures, then it is important that this epistemic nature of the discipline is present in the formulations of the curriculum (Wheelahan, 2007, p. 648).
The issue of vertical knowledge in curricula can be viewed from several perspectives. Young and Muller are among those that argue for the inclusion of vertical knowledge in curricula. This is based on ethical concerns, that no group of students should be denied this kind of knowledge, defining it as 'powerful knowledge' (Young, 2008). Beck is more hesitant about this idea as he argues that there may be students who, due to their limited exposure to vertical knowledge at home, become disadvantaged at school in relation to students who are exposed to such knowledge continuously as they grow up (Beck, 2013, p. 187). This line of argument is given support by research indicating that students can show some resistance when vertical knowledge discourses are present in education because the students perceive no obvious connections to their life worlds (Doherty, 2015, p. 720). The inclusion of horizontal knowledge is often made with arguments related to issues of meaning and also often result in the removal of vertical aspects. This indicates that the inclusion of vertical aspects into curricula makes the teaching practice central, if all students should have the possibility to get access to these vertical aspects. This has been exemplified by McCabe, showing that students can be reluctant to engage with vertical knowledge unless a teacher guides them in the process (McCabe, 2017, pp. 428-429).

From the discipline to the curriculum
Recontextualization is a concept that is used to understand the process when knowledge from various fields of knowledge production is brought into school curricula. The idea behind the concept of recontextualization is that the relationship between a certain body of knowledge and the formulations in the curriculum document rarely is a one-to-one relationship. Instead, the aspects from the fields of knowledge production are transformed into formulations in the curriculum through a recontextualization process (Bernstein, 2000, pp. 33-34).
Recontextualization can take place in two separate fields. One of these is the official recontextualization field-the field in focus in this article. It should be understood as the governmental agency responsible for the formulation of curriculum documents. The other field is the pedagogical recontextualization field, which consists of a number of actors, textbook authors, lecturers of teacher education and people who engage in public debates about educational issues. In relation to these fields, recontextualization agents influence the processes within each of them (Bernstein, 2000, pp. 33-34).
A second concept that is used in the article is field of knowledge production. A field of knowledge production should be understood as the origin of the contents of a curriculum. This could be, for example, the scientific disciplines of history, political science or chemistry. Although, as Muller (2009) has discussed, the scientific status of history is contested, the relevance of the historical discipline in this study is the specialized knowledge that it has developed over time. Knowledge produced in these fields of knowledge production are, through the process of recontextualization, incorporated into the curriculum. As mentioned, the aspects of a field of knowledge production are seldom transferred in a one-to-one relationship, which means that certain aspects are selected from a field of knowledge production and relocated into a curriculum (Bernstein, 2000, pp. 33-34).
A school subject can be composed of elements from one or several fields of knowledge production. The number of fields of knowledge production that are drawn upon in the recontextualization process can affect the degree of classification of the school subject. Classification refers to the strength of the boundaries between different curriculum categories, and in the context of this study, these categories are school subjects. Classification then means how delineated a subject is in relation to other subjects (Bernstein, 2000, pp. 5-6).
Formulations in a curriculum can leave more or less room for teachers to make decisions about how they address a certain element from the curriculum in their teaching. A third concept, framing, is used to further the understanding of the distribution of control between curriculum and teachers. A large degree of framing implies that the formulations in the curriculum leave little room for teachers to make decisions because the curriculum clearly stipulates how a certain feature of the curriculum is meant to be addressed. A lesser degree of framing means that the formulations in the curriculum are not as decisive in relation to a certain feature. This leaves more room for decisions to be made by the teacher (Bernstein, 2000, pp. 11-12). Changes of curricula are often discussed in positive or negative terms. Whether a change in the degree of classification or framing is perceived as positive or negative is dependent on how one evaluates its consequences.
On the issue of vertical and horizontal knowledge structures, Bernstein argues that what characterizes these structures are both historically and culturally dependent (Bernstein, 2000, p. 29). I view the concepts mentioned in this section in the same way. The strength of classification, for example, can be viewed as strong in one context and weak in another.
In the following section, I will relate this discussion on different types of knowledge discourses to history education and school history curricula. To relate the issue of specialized vertical knowledge to a process of recontextualization, it is necessary to sort out what aspects from a field of knowledge production are chosen to be included into a formal history curriculum. The fact that the discipline of history is regarded as a vertical knowledge structure does not automatically mean that this verticality is maintained through a process of recontextualization. It is thus not self-evident that the verticality is also present in a certain history curriculum. To discuss this, I will address the relationship between history education research and the historical discipline.

Verticality and history education
In history education research, there are those who argue that the ways in which the historical discipline relate to history are counter-intuitive in ways similar to vertical knowledge. Sam Wineburg has described the disciplinary way of thinking historically as an unnatural act. This formulation of 'unnatural' is used to express that the discipline's way of approaching the past is knowledge that in most cases must be learnt through being schooled in formal history education (Wineburg, 2001, p. 7). Similarly, and much in line with Muller's argument, is the discussion where Peter Lee differentiates between the substantive knowledge of history on the one hand and metahistorical concepts on the other. Metahistorical aspects address the epistemic nature of the history-how the historical knowledge is produced and structured in the historical discipline. Metahistorical concepts include evidence, continuity and change, cause and consequence, historical empathy and historical significance (Lee & Ashby, 2000;pp. 199-201;Lee, 2004;pp. 140-143). In history education research, these concepts are labelled 'second-order concepts'. As a way to establish the presence of these concepts in the historical discipline, Peter Seixas studied the work of several historians, and he argues that the aforementioned second-order concepts are present and greatly significant in these works (Seixas, 2015a). However, in order to make teaching of second-order concepts constructive, it is necessary that teachers have the necessary competencies to teach those (Wilson & Wineburg, 2001).
Following this line of discussion about the nature and the eventual power of vertical knowledge discourses, the inclusion or exclusion of second-order concepts in history education is important. If these kinds of second-order concepts (derived as they are from the historical discipline) are not brought into a recontextualization process, it would mean that the epistemological aspects of the historical discipline-its verticality-are not addressed in the curriculum. The inclusion of the verticality demands not only that the disciplinary aspects-the metahistorical, second-order concepts-are brought into the recontextualization process but also that these aspects are addressed so that the verticality is present in the curriculum documents in a way that makes them accessible for teachers, and by extension, also for the students.
How vertical aspects are handled in processes of recontextualization has been examined in relation to several subjects. Macdonald et al. examine the recontextualization in health education and identify a misrepresentation of the knowledge brought in from the field of knowledge production and recontextualized into an educational context (Macdonald, Hunter, & Tinning, 2007, p. 115). This would mean that there is a risk that formulations in curriculum documents lack certain important aspects present in the original body of knowledge. McPhail uses the concept of recontextualization to examine the use of constructivism as a learning theory in curriculum documents. He argues that there is a danger that constructivism, when recontextualized into curriculum documents, becomes a theory of everything and fails to address the epistemological aspects of the concept (McPhail, 2016, p. 304). Thobani examines the recontextualization process of the school subject of religious studies in England. He argues that the diversity with how Islam is addressed in academia is lost in the school curriculum document, where it is presented as a monolithic entity (Thobani, 20115, p. 42). These studies indicate that aspects from the fields of knowledge production are not always carried through the recontextualization process and brought into curriculum documents. In the following sections, I will address the recontextualization process in more detail.

Tools for formulation of level descriptors
The content of the Swedish subject curricula is supposed to be assessed by the teachers, who are to relate their assessments to the formulated level descriptors. One can think of a variety of origins for the criteria directing the formulation of level descriptors. A generic take on different levels of knowledge implies that there are commonalities between being proficient in, for example, history and chemistry. Bloom's taxonomy (Bloom, 1956), the revised version of Bloom's taxonomy (Anderson et al., 2001) and the Solo Taxonomy (Biggs, 1982) are examples of such generic approaches to criteria for definitions of level descriptors.
There are also subject-specific approaches to defining levels of proficiency. A number of models describing such levels related to the subject of history were developed in relation to the two projects, SHP and CHATA, conducted within the English school system. The aim of these projects was to include aspects of the historical discipline, the aforementioned second-order concepts, into history education. Empirical research conducted in relation to these projects resulted in a number of progression models, each addressing one second-order concept. The models were not constructed to serve as level descriptors, but they are useful when discussing qualitative levels in history education. These models address how students perceive the nature of historical accounts (Ashby, 2011), how they view the ways in which evidence is constructed (Lee & Shemilt, 2003), historical empathy (Lee & Shemilt, 2011) and how students understand the conceptual pair of cause and consequence (Lee & Shemilt, 2009). One common feature in all these models is the importance of an understanding of the epistemic nature of historical knowledge. Empirical research on how students establish connections between the past and the present show that students on the lower levels do not motivate opinions on present situations or connect them with history. On higher levels, they can connect a present event with history as well as connect various events in the past to explanations of present situations (Duquette, 20155, pp. 55-56;Rosenlund, 2016;p. 169 Sandahl, 2015;p. 9).

Methodological consideration
The empirical material in the study consists of the curriculum for the first history course for the academic tracks in the Swedish upper-secondary school, labelled History 1b (Swedish National Agency for education, 2012). This curriculum, as all subject curricula in Sweden, consists of three sections (see Table 1, below): (a) the aims of the subject (p. 2), (b) the core content that has to be addressed in the course (p. 9) and (c) the knowledge requirements that should be used by teachers Table 1. Three sections of the Swedish history curriculum.

Aims
Core content Knowledge Requirements The purpose of history education is concretized in five items describing what education should help students develop. One item addresses 'knowledge' and four address 'abilities'.
What teaching in history should address is described in six items. These items address substantive processes, second-order concepts, types of sources and different uses of history.
Descriptions of three levels of proficiency, related to the five aims.
when they give students their grades (p. 10-11). To address the issue of internal consistency in this history curriculum, the method of content analysis was used. This method was suitable for this study as it provides a transparent framework for analysing the content and meaning that can be inferred from texts (Krippendorff, 2004, p. 44), a curriculum document in this case. Such a framework was especially necessary in this study as it includes a comparative approach. The coding of the curriculum text was directed at three aspects: the first addressed the fields of knowledge production that the content of the three sections curriculum can be related to, the second addressed the degree of framing in the three sections of the curriculum and the third aspect addressed the degree of classification of the curriculum sections. To increase the reliability of the analysis, I used the same deductive and inductive strategies when analysing the three sections of the curriculum. A deductive strategy was used initially as I was using a predefined set of theoretical concepts in order to examine the fields of knowledge production. Each of the items in the three parts of the curriculum was examined in relation to the second-order concepts formulated within history education research. In the coding process, the items that could be related to these secondorder concepts were categorized as belonging to a vertical knowledge discourse. An inductive strategy was applied in the cases where items in the curriculum could not be related to the secondorder concepts defined in history education research. Then other fields of knowledge production were considered as possible fields of knowledge production. This was the case for two items in the aims, two items in the core content as well as the main theoretical framework underpinning the level descriptors in the knowledge requirements section. The classification and framing of the curriculum's three sections are presented as tentative discussions. To increase the transparency of this part of the study, it has been my intention to provide a sufficient amount of examples from the curriculum document as evidence for my claims. The use of content analysis helped both to enhance the transparency of the interpretation process and to increase the understanding for the three aspects examined: fields of knowledge production, framing and classification. By comparing the results from the analysis of each of the three sections in the curriculum, it was possible to draw conclusions regarding the internal consistency in the curriculum.

The Swedish history curriculum
Sweden shifted from a norm-referenced curriculum to a standards-based curriculum in 1994, which was retained in the 2011 curriculum reform. Formulated by the Swedish Ministry of Education, two directives of the reform were as follows: to clarify the level descriptors, whose label was changed from grading criteria to knowledge requirements, and to rely on the definition of knowledge that was formulated in the curriculum of 1994 (Swedish national agency for education, 2015). This view of knowledge is based on the assumption that all knowledge consists of four aspects that are assumed to presuppose and interact with each other: facts, understanding, skills, familiarity and accumulated experience. There was a shifting composition of recontextualization agents that participated in the various parts of the 2011 reform. History education researchers, history teachers and officials at the National Agency for Education together formulated the sections addressing aims and the core content. The history education researchers and teachers were not involved in the formulation of the knowledge requirements-a work conducted solely by officials at the Agency for Education (Swedish national agency for education, 2015, pp. 45-46). The reform of 2011 resulted in a curriculum where two sections were preserved from the previous curriculum: the aims and the level descriptors. A change in the new curriculum was the inclusion of a section labelled core content. This section describes aspects that are to be addressed in each course. In relation to assessment, it can be noted that in 2013 (two years after the curriculum was implemented), a mandatory national test in history was introduced for students in year nine and a voluntary test was introduced for history students in upper secondary school. In the following, the three sections in the curriculum, aims, core content and knowledge requirements (see Table 1) will be described and examined with regard to what fields of knowledge production that can be identified in them and their degree of classification, framing and verticality. The aims and the core content will be addressed in one section, while the knowledge requirements will be addressed in the subsequent section.

Aims and core content
As mentioned above, the purpose of history education is concretized in five aims, and they are formulated as follows.
Teaching in the subject of history should give students opportunities to develop the following: (1) Knowledge of time periods, processes of change, events and persons on the basis of different interpretations and perspectives.
(2) The ability to use a historical frame of reference to understand the present and to provide perspective on the future. (3) The ability to use different historical theories and concepts to formulate, investigate, explain and draw conclusions about historical issues from different perspectives. (4) The ability to search for, examine, interpret and assess sources using source-critical methods and to present the results using various forms of expression. (5) The ability to investigate, explain and assess the use of history in different contexts and during different time periods.
These five, rather abstract, aims do not contain any specified historical content. As mentioned in the previous section, the content of history education is described in a new section-core content-introduced in the curriculum after the 2011 reform. In the 1994 curriculum, no specification for what content should be addressed in history education existed. Each of the six items in the new core content section can be related to one of the aims, which means that the core content does not solely address a traditional historical content. There are items in the core content section that are related to historical concepts, to the use of historical sources and to how history is used in the present. One example of an item in the core content that is related to uses of history is: 'How individuals and groups have used history in everyday life, societal life and politics'.

Fields of knowledge production-aims and core content
When it comes to what fields of knowledge production are present in the two sections aims and core content, I would argue that there are distinct influences from two fields of knowledge production. The historical discipline as a field of knowledge production is visible in the first, third and fourth aims. The first aim addresses time periods and processes of change, the third aim addresses historical theories and concepts and the fourth aim addresses interpretations of historical sources. These three aims can all be related to the discipline of history (Jordanova, 2006, pp. 35-41, 100-102) and have been present in Swedish history curricula since the middle of the 20th century (Rosenlund, 2015). There are also elements in the aims that can be regarded as originating from fields of knowledge production other than the traditional discipline of history, and these are visible in the second and the fifth aim, where historical knowledge is connected with the present and the future. In the following, I refer to these aims as 'use of history'. The second aim can serve as an example of this, where it is stated that students should develop their 'ability to use a historical frame of reference to understand the present and to provide perspective on the future'. The reason that this is interpreted as not coming from the historical discipline is that it explicitly states that a link between the past, the present and the future is to be addressed in history education. This is an activity that might not stem from the traditional historical discipline, even though there are historians that focus their interest around this connection (Green, 2012;Staley, 2007). However, the inclusion of these temporal connections is not met with enthusiasm from all areas of the historical discipline (Bonneuil, 2009). One might view the inclusion of these temporal connections into the curriculum as an influence from a field of production separated from, but still linked to, the discipline of history. This field is often referred to as 'history education research', and it differs from the historical discipline in that it is focused on the transmission and use of historical knowledge rather than on extending and deepening knowledge about the past. Within this educational field of knowledge production, the issue of temporal connections between history, the present and the future is of some importance (Karlsson, 2014;Körber, 2015;Rüsen, 2004;Seixas, 2015b). The contents of these two aims, addressing temporal connections and the use of history, have been discussed within history didactics in Sweden since the 1990s (Karlsson, 2009, p. 47-48). More recently, Nordgren suggested that use of history should be regarded as a third aim of history education, to complement the two more established aims: knowledge about the past and historical thinking (Nordgren, 2016). As the section describing what core content that should be addressed in the course consists of six paragraphs, and each of these can be related to one of the aims of the subject, one could say that the core content prescribes the content with which the aims should be addressed. Regarding possible fields of knowledge production, a similar discussion can be made regarding the paragraphs in the core content as was the case with the aims. There are five items in the core content that can be related to the discipline of history, addressing historical knowledge about processes and classifications, historical concepts and interpretations of sources. In addition, one paragraph addresses the use of history that can be related to the field of knowledge production of 'history education research'.

Classification
Looking at the aims of the subject, there are three of the five aims (aims 1, 3 and 4) that I would argue provide the curriculum with classificatory strength. This is, as discussed in the section above, because these aims are formulated in ways that make it evident that they belong to the sphere of historical knowledge. The two remaining aims, related to the use of history, might be considered to provide the curriculum document with some classificatory weakness. This is because they can be perceived to move the focus of history teaching towards the present and/or the future, something that may make the history subject's borders to other subjects less distinct (Sandahl, 2015, pp. 9-11).
Similar reflections can be made in relation to the section describing the core content of the course. The formulations related to the historical discipline can be argued to bring classificatory strength to the document. The one formulation in the core content related to use of history, however, might be perceived as not being clearly related to the historical sphere, as it connects history with the present.

Framing
The issue of framing addresses the distribution of control of the content of history education: What decisions are made by the curriculum designers and what decisions are distributed to the teachers? The aims that are formulated in the Swedish history curriculum can be argued to be loosely framed, as the aims do not specify what content should be addressed in the classroom. One example is one of the aims regarding uses of history-that students should develop their 'ability to use a historical frame of reference to understand the present and to provide perspective on the future'. In this, as in the other aims, there are no explicit formulations about what content the various abilities should be developed with.
However, the perception of a loosely framed curriculum changes when the section describing what core content teachers should address is included in the examination. When the core content is related to the aims, it becomes evident that the aims related to historical knowledge should be handled in relation to the following item in the core content: 'Industrialisation and democratisation during the 19th and 20th centuries in Sweden and globally. . .'. The content of education is more strongly framed here, especially when compared with the curriculum from 1994 where there was no description of what historical content to address. However, the description of a core content in the curriculum leaves room for decisions to be made about what to address in the classroom to the individual teacher. For example, there is room to choose various aspects of a democratization process if one is to focus more on the liberalization of society in the 19th century or more on the fight for universal suffrage in the beginning of the 20th century. However, the use of very detailed learning outcomes in some countries (Bertram, 2016) indicates that there are curricula more strongly framed than the Swedish curriculum.

Verticality
When it comes to the type of knowledge present in the Swedish subject curricula following the reform in 2011, Sundberg and Wahlström argue that it can be described as 'representational knowledge' and that the curricula promote a view of knowledge as a product (2012, p. 351). The following discussion will provide a complementary interpretation based on an examination of the history curriculum.
The first aim for history education in the curriculum addresses a historical content, and following Muller's definitions of verticality, I would argue that the formulations indicate that we are here dealing with the product of the historical discipline (Muller, 2014, pp. 261-263). The third and the fourth aim, which address theories, concepts and historical methods, can both be regarded as carrying verticality from the historical discipline into the history curriculum. This is because they can be connected to procedural and epistemological aspects of the historical discipline. These are aspects that are labelled as second-order concepts within history education research, where they are ascribed a certain importance because of their transferability and that they seem to be difficult to understand, if not learned in formal schooling.
The two aims in the curriculum addressing uses of history (2 and 5), I would argue, contain aspects of what Muller terms 'inferential knowledge' and are used to make connections between the declarative products of a discipline. The formulations of the aims indicate that the competencies they specify are transferrable. There is also research indicating similarities between the two aims of historical orientation and second-order concepts. These studies show that students have difficulties with uses of history that are similar to those when students approach second-order concepts (Duquette, 2015;p. 51;Rosenlund, 2016;pp. 168-169;Seixas, Peck, & Poyntz, 2011;p. 47). These difficulties include the fact that some students seem to have no previous experience of engaging in historical orientation, and although they are explicitly encouraged to do so in these studies, they do not seem to be able to. Taken together, these characteristics make 'uses of history' similar to 'second-order concepts', and it is thus possible that they can be regarded as carrying verticality into the curriculum.

Level descriptors in the knowledge requirements
The third area of interest in this study is the level descriptors in the knowledge requirements. The knowledge requirements do not target the specific content mentioned in the core content but rather are related to the aims. The connection between the three parts of the curriculum can be concretized with this example. The first aim states that the students should develop 'Knowledge of time periods, processes of change. . .'. In the core content, this is concretized by the formulation, 'democratisation during the 19th and 20th centuries'. In the knowledge requirements, a formulation states that students should 'give an account of processes of change', and the level descriptors describe three levels of quality with which students can give such an account. In the following, the level descriptors in the knowledge requirement section will be discussed in more detail.
The curriculum reform in 2011 resulted in the heading of the level descriptors being changed from 'grading criteria' to 'knowledge requirements'. Sundberg and Wahlström claim that the formulations of level descriptors are new to the 2011 curriculum (Sundberg & Wahlström, 2012, pp. 350-351). However, the presence of level descriptors is one sign of continuity between the two curricula. Both the 1994 curriculum and the curriculum from 2011 contain descriptors for three levels (Rosenlund, 2016, pp. 69-70, 122). In the curriculum from 1994, verbs similar to those found in the two versions of Bloom's Taxonomy are used to differentiate three levels of student achievement. In the 2011 curriculum, a different principle is used to differentiate between levels, as adverbs are used instead of verbs (Rosenlund, 2016, p. 122). The adverbs are present in the level descriptors of all subjects in the Swedish primary-and upper-secondary school. The following are examples of how adverbs are used in the history curricula to differentiate among levels E, C and A (regarding one aspect of historical methods).
Level E: Students can with some certainty search for, examine and interpret source material to answer questions about historical processes and also make simple reflections on the relevance of the material.
Level C: Students can with some certainty search for, examine and interpret source material to answer questions about historical processes and also make well-grounded reflections on the relevance of the material.
Level A: Students can with certainty search for, examine and interpret source material to answer questions about historical processes, and also make well-grounded and balanced reflections on the relevance of the material.
As seen in the example, the same adverb phrase, 'with certainty', is used in the level descriptors to describe student performance when it comes to the searching, examining and interpreting of historical sources. This means that the same cognitive process is assessed on these levels and that it is the quality with which the student performs this process that is evaluated. When it comes to the next part of the paragraph, addressing students' reflections on the relevance of sources, the adverb 'simple' is used to describe Level E and the adverb phrase 'well-grounded' is used to describe Level C. Level A contains the adverb phrase 'with certainty' to describe performance related to the first part of the paragraph and 'well-grounded and balanced' to describe reflections on relevance. This way to describe levels of student performance is used throughout the knowledge requirements, with three exceptions where other principles for the level descriptors are used. One of these exceptions will now be examined in detail.
In the first of these exceptions, the verb compare is used to define the difference between two levels of performance. The second example of an exception is where a quantitative aspect is used to differentiate between levels regarding the ability to use historical theories and concepts. The third example of an exception from the rule of adverb-driven descriptors relates to historical methods (see the example above) and will be addressed more in detail.
On Level E, students should make assessments based on source-critical criteria, while on Level C, students should make evaluations based on source-critical methods. Here, the progression seems to go from criteria to methods. I make the interpretation that criteria refers to the use of individual criteria such as 'bias' and 'dependency', while methods refers to a more consistent use of source-critical thinking. There is an additional aspect in this paragraph that indicates the difference between levels.
On the E level, students are required to relate the result of a source-critical evaluation to an interpretation of the same source(s). However, on levels C and A, the students are required to relate their evaluation to different possible interpretations of the source(s). Here, the progression seems to lie in moving from evaluating one interpretation of a source material to evaluating different interpretations of a source material. The difference between the levels is related to the nature of the historical interpretations, where the assessment of several interpretations of a given material is regarded as more complex than the evaluation of a single interpretation of the same source material.

Fields of knowledge production
As discussed above, the main principle for formulation of level descriptions is the use of adverbs, and it is not possible to relate these adverbs to any relevant field of knowledge production. According to the Swedish National Agency for Education's own evaluation of the 2011 reform, the use of adverbs was a way to comply with the view on knowledge formulated in the general part of the curriculum. The intended use of the revised version of Bloom's Taxonomy was abandoned due to this, and the agency instead chose the adverbs as the principle for level descriptions. There are no indications in the evaluation as to any source of the inspiration for the use of this principle (Swedish national agency for education, 2015, pp. 55-60).
The above-mentioned exception from the adverb-driven principle for level descriptors is related to historical methods. In this case, the principle of adverbs is accompanied by what can be interpreted as a disciplinary approach to level descriptions. In this case, the descriptors are possible to trace to discussions within the discipline of history where the complexity of interpretation often is mentioned as a sign of an increase in quality (Berkhofer, 2008;pp. 54-55;Jordanova, 2006;pp. 160-161). It also possible to see the history education research as a field of knowledge production in this case. The level descriptors that are used in relation to historical methods share many similarities with ideas on progression of the second-order concept of evidence (Lee & Shemilt, 2003).
To conclude, one major principle governs the way the level descriptors are formulated. This principle is based on the use of adverbs as characteristics for the three levels. What field of knowledge production is drawn upon in this case is uncertain, but one reason for choosing the adverbs is that the agency did not want a taxonomic, hierarchical principle, with verbs defining the level descriptors. One factor that can help explain this can be found in the shifting composition of recontextualization agents that participated in the curriculum reform, which was presented above. The sections of the curriculum that include examples of vertical knowledge discourses were formulated in collaboration with history education researchers, who brought knowledge about the vertical aspects of history education research and the historical discipline into the process of formulating the curriculum (Swedish national agency for education, p. 45, and Ledman, 2014;p. 163-164). In contrast, the level descriptors-which cannot be understood as representing a vertical knowledge discourse-were formulated by officials at the agency, which did not interact with the history education researchers who had formulated the other parts of the curriculum (Swedish national agency for education, 2015, p. 46).

Classification
The degree of classification of the knowledge requirements varies between different parts in this section of the curriculum. The majority of the level descriptors are formulated along the same method as with all other subjects in upper-secondary school in Sweden-with adverbs, and they are the same adverbs that are present in all subjects. This means that the adverbs that are used to define the three levels are the same in history as they are in chemistry or geography. The fact that there is nothing specifically historic with the general principle for differentiating between levels of competency gives this part of the curriculum a weak classification. There is one part of the knowledge requirements, that of historical methods, which is differently classified. As discussed above, the formulation in this part of the knowledge requirements can be related to the discipline of history. This gives this part of the knowledge requirements a stronger classification than the other parts.

Framing
In the section labelled 'knowledge requirements', the distribution of control over definitions of level descriptors can be examined. Teachers are to decide on which of the three levels the students have performed. However, the fact that there are descriptions for all three levels of assessment can be construed as giving little room for teacher autonomy when it comes to assessing the level of students' knowledge in history, and thus, the framing of assessment could be described as very strong in the Swedish curriculum. This is an interpretation brought forward by Ledman (2014, p. 175). The argument for such a view is that the formulation of descriptors for three assessment levels leaves the teachers with little room for control over assessment. Such an interpretation is based on the assumption that the type of knowledge required to achieve a certain grade is clearly stipulated. However, it is worth considering to what extent generic formulations (like the adverbs in the Swedish curriculum) that do not originate from an established field of knowledge production can impose control over teachers' assessment. It may be the case that formulations like 'in a balanced way' are perceived by teachers as vague, and such vagueness may lessen the degree of framing of the curriculum and widen the room for interpretation, thus increasing the teachers' degree of control.

Verticality
The fact that the assessment-related level descriptors are incorporated not from the discipline of history, history education research or educational psychology raises questions about how the verticality of the historical discipline is preserved throughout the curriculum. It is possible that in order to preserve the vertical qualities of a subject through a recontextualization process, it is necessary not only to have the various aspects of a subject in mind but also how different levels of quality are expressed within these fields of knowledge production. In the Swedish case, this is only done in relation to historical methods, and thus the verticality present throughout the aims and the core content is almost invisible in the level descriptors.

Conclusions
Regarding the internal consistency regarding vertical aspects, I argue that an asymmetrical relationship exists between the different parts of the Swedish history curriculum. Many of the aims and much of the core content consist of items that share the vertical characteristics of the historical discipline. However, the majority of the formulations in the knowledge requirements communicate a different knowledge discourse, as the level descriptors are not incorporated from the historical discipline, history education research or any other scientific discipline. This inconsistency of the communicated knowledge discourse in the Swedish curriculum has several possible consequences. One of these is that the reliability of teachers' assessment can be weakened. There are two main reasons why this is likely. First, as the same level descriptors are used for all subjects, this can be construed by teachers as evidence that established models of level descriptors (subject-specific or generic) for differentiating between student performances should not be used. This leaves teachers with little assistance in the assessment of their students. A second, related, reason is that the adverbs actually impose a low degree of framing with regard to teachers' assessment of students' knowledge. This makes it likely that teachers will comprehend the meaning of the level descriptors in a different way. This is one plausible factor behind the problems with inter-rater reliability in the Swedish educational system that the Swedish Schools Inspectorate has highlighted (Swedish Schools Inspectorate, 2017, p. 5). Consequently, teachers to some extent may make different interpretations of what constitutes high proficiency in relation to the five aims of the history curriculum. In addition, I would argue that different interpretations of what expressions of high proficiency look like can also lead to different approaches to teaching history. This is because a teacher's idea of what characterizes high proficiency is likely to have an impact on how specific lessons are designed.
A second consequence of the inconsistency is that the usefulness of the vertical knowledge discourse risks being diminished. In the Swedish curriculum, when second-order concepts like 'evidence' (labelled interpretation of sources in the Swedish curriculum) are incorporated into the curriculum and the students' performance is assessed using generic adverb phrases like 'certainty' and 'with certainty', the verticality is likely to be affected. Since subject-specific descriptions of qualitative levels are not prescribed in the curriculum, the likelihood that the full potential of the vertical second-order concepts is made available for the students may be reduced. It will thus be difficult for teachers to create a history education that helps students acquire what is described in the aims of the history subject. One way to come to terms with the inconsistency, and thus increase the possibility to retain the verticality through the curriculum, is to adopt an evidencecentred design (Ercikan & Seixas, 2011) and formulate level descriptors based on relevant research in history or history education research. A higher degree of consistency would increase the probability that the teachers provide access for the students to the full potential of the vertical knowledge discourses included in the curriculum. There is, however, one possible obstacle to this: it concerns the extent to which teachers themselves have command over and can teach the vertical knowledge and whether there are educational resources to support teaching and learning. There are indications that, in a Swedish context, it is not always evident that this is the case (Jarhall, 2012;Rosenlund, 2016;Swedish Schools Inspectorate, 2015). This begs the question whether inclusions of vertical knowledge discourses may be no more than a Potemkin village: looks good at a first glance, but has no real impact in the classrooms.
The exclusion of history education researchers from the process of formulating the knowledge requirements can be one factor that helps explain the inconsistency in the history curriculum. This is because the subject-specific knowledge held by these history education researchers was not accessible for the officials that formulated the level descriptors. The Swedish example presented in this article indicates that one way to achieve a higher degree of internal consistency in curriculum documents is to have agents from the same fields of knowledge production represented in the formulation of all sections of the curriculum.
That there are descriptors for three levels in the Swedish subject curricula has been construed as evidence that it is strongly framed and leaves little power in the hands of the teachers. Supported by the study presented in this article, I would like to provide a different perspective of the Swedish subject curricula. It can be argued that the vagueness inherent in the adverbs in the level descriptors can give teachers increased control over how to assess and grade their students. It is possible, however, that the degree of framing is strengthened by the use of large-scale testing implemented in history some years after the curriculum reform. The definitions of the different levels that are present in the tests can entail that the control over how to assess has been transferred-not from the curriculum to teachers, but from the curriculum to those constructing the tests. A process where large-scale testing influences how teachers plan and execute education has been identified in many settings (Grant, 2015, Löfström, Virta, & Van Den Berg, 2010. Although there has been no research regarding the subject of history in Sweden, there are indications that this process is also prevalent in Sweden (Arensmeier et al., 2016).
As mentioned above, Beck directs our attention to possible disadvantages for certain groups of students when vertical knowledge discourses are included in the curriculum. If Beck's assumption holds, equality would be endangered, unless measures were taken to increase the possibility for all students to learn this vertical knowledge by ensuring that history teachers have the competences necessary to teach the second-order concepts of history education. It is possible that a low degree of framing in the knowledge requirements can be used by teachers to counteract and reduce the eventual negative impact that vertical knowledge discourses may have on disadvantaged groups of students. A higher degree of framing would reduce the room for teachers to mediate between the curriculum and students' knowledge.
Curriculum construction requires that attention be paid to many aspects that are related to each other in complex ways. Even if there is research that can help inform recontextualization agents, some decisions are bound to be taken on ideological (educational as well as political) grounds. Based on the findings presented in this article, I would like to recommend curriculum designers to (1) include researchers who are familiar with the subjects in the curriculum that are affected by the reform, and (2) at the outset of a reform process, articulate the desirable inclusion of vertical knowledge and the desired degrees of framing and classification. Such clarifications would make it possible to discuss and address some of the possible consequences beforehand.