IMPLEMENTING GENRE PEDAGOGY IN CONTENT INSTRUCTION: LESSONS FROM SWEDEN

Drawing upon studies conducted in Sweden, this article discusses possibilities and limits of implementing genre pedagogy in content instruction. The wider educational concern is how knowledge of genre and language can be used to promote a deeper engagement with content knowledge. The linguistic theory underpinning genre pedagogy and the pedagogic-practical teaching/learning cycle is explained. Then, two empirical studies of genre-based teaching in Geography in Grade 6 are reviewed, with a particular focus on the texts used as models for the students’ own writing. The studies show two contrasting sides of genre-based intervention: one in which generic structures and other features of texts are used productively to engage with content knowledge and one in which attention to generic structure and logical connections comes at the expense of the negotiation of content knowledge. The article concludes with recommendations for implementing genre pedagogy.


Introduction
The aim of this paper is to discuss limits and possibilities of implementing genre pedagogy in content instruction. The overarching educational concern to be addressed is how knowledge about language can promote a deeper engagement with instructional content. After a general introduction to the concept of genre pedagogy, the discussion will draw upon studies and interventions conducted in Sweden pertaining to the teaching of Geography in Grade 6.
Genre pedagogy, or genre-based instruction, originated in Australia during the 1980's in order to enhance the prospects of educational achievement among marginalized groups (Rothery, 1996;Feez, 2002;Rose & Martin, 2012). This pedagogy was also a reaction to prevalent progressivist or constructivist approaches, which were accused of obfuscating what needed to be learnt by using unclear criteria, vague boundaries between disciplinary domains and noninterventionist teaching approaches. Genre pedagogy was promoted as a subversive visible pedagogy which, based on Bernstein's sociology of education (Bernstein, 1990(Bernstein, /2003(Bernstein, , 2000 and Systemic-Functional Linguistics (SFL) (Halliday & Mathiessen, 2014), sought to make implicit demands for school achievement explicit through the teaching of linguistically defined genres (Rose & Martin, 2012;Martin, 1999). In Sweden, genre pedagogy has gained considerable traction in the last decade in promoting second language learners' prospects of developing content knowledge and linguistic skills simultaneously (Walldén, 2019a). The overall instructional aim is to combine high levels of support with high levels of cognitive challenge (Mariani, 1997;Gibbons, 2006). In this paper, I will explain the theory behind genre pedagogy and, drawing upon two classroom studies, discuss examples of successful and less successful implementations.

Genre theory
To give a comprehensive view of the linguistic theory underpinning genre pedagogy is beyond the scope of this paper. However, I will draw attention to salient points relevant to the following discussion.
According to Martin's genre theory (Martin & Rose, 2008;Martin, 1992;2001), genre is defined as a staged, goal-oriented social process. Crucially, genres are understood in terms of internal features rather than external ones pertaining to the rhetorical situation (Feez, 2002;Paltridge, 2014). Therefore, the names of the genres reflect broad communicative purposes, such as arguments and explanations. The staged part of Martin's definition concerns the generic structures. For example, discussion (a subgenre of argument) can be expected to adhere to the following structure: introduction of the issue, review of competing sides and a concluding position statement (Christie & Derewianka, 2010). Similarly, a factorial explanation (a subgenre of explanation) starts with the introduction of a phenomenon and moves on through the explanation of different factors leading up to said phenomenon (Martin & Rose, 2008).
Generic structures such as these can be used for dealing productively with content knowledge in different disciplines. However, Martin's theory has been criticized for locking communicative goals to specific generic structures and thus limiting the possibilities of expression (Holmberg, 2012;Hasan, 1995Hasan, /2016Freedman, 1994;Watkins, 1999). It certainly seems possible to advance an argument by drawing upon non-argumentative generic structures, such as explanations or narratives. Such concerns, which are also frequently expressed in Sweden (Liberg, Wiksten Folkeryd & af Geijerstam, 2012;Liberg, 2008;Hertzberg, 2006), seem to overlook the fact that Martin describes the relationship between generic structure and communicative purpose as probabilistic rather than deterministic (Martin, 2001). Moreover, Martin uses the term contextual metaphor to describe precisely how a certain generic structure can be used to fulfill a less typical communicative purpose (Martin & Rose, 2008).
It is also crucial to note that, according to Martin's theory, generic structure is only a part of what constitutes a genre. Using a genre successfully also involves linguistic choices relating to the register variables of field, tenor and mode (Rose & Martin, 2012;Martin & Rose, 2008;Martin, 2001). Field concerns the experiential content of the discourse as well as logical connections. In disciplinary writing, students are often required to handle technical and abstract vocabulary, and to describe relevant processes and activities pertaining to the disciplinary domain explored. The linguistic category of grammatical metaphor is an important resource in accomplishing this. It involves experiential metaphors, which transform processes (such as "to pollute") into things ("pollution") or qualities ("polluted"), and logical metaphors which expresses logical connections as things (e.g. "result", "consequence") or prepositional phrases (e.g. "due to") instead of as conjunctions (e.g. "because"). Grammatical metaphors are necessary for producing the kind of heavily nominalized discourse integral to engaging with knowledge in different disciplines (Martin, 1990(Martin, /1993Martin, 2009Martin, , 2013. Disciplinary literacy often requires the students to develop an expert voice by communicating knowledge in an authoritative fashion. While employing abstract and technical language and formulating logical relations are important, there are also choices related to the register variable of tenor: how the text reflects and construes the relationship between writer and recipient. For example, an expert voice likely involves the use of declaratives rather than questions and exclamations, and the use of objective modality ("it seems necessary to") or passive voice ("X is recommended to") rather than more subjective wordings ("I think it's important to", "X should") (Schleppegrell, 2004;Martin & White, 2005). Another convention in more distanced for of wiring is restricted use of personal pronouns in first and second person. Finally, a measured use of evaluative language can be expected to either promote or undermine the perspective discussed in a text (e.g. "a valid objection", "a far-fetched conclusion") or to construct a moral position in an explanation ("a dire consequence") (Christie & Derewianka, 2010: Martin & White, 2005. Of course, the grade in which the instructions takes place must be considered: what constitutes a weak expert voice in later years of schooling might pass as entirely appropriate in earlier years (Christie & Derewianka, 2010). The important point is that the teaching offers opportunities for all students to expand their meaningmaking capabilities in ways which are valued highly in the assessment of disciplinary writing.
As for mode, disciplinary literacy often requires that students regulate the information flow in their text in a predictable and planned manner. In discussions, marked textual themes can be used to guide the reader through the staging of the text: "On one hand…, "On the other hand …", "In conclusion …" (Martin & Rose, 2008;Christie & Derewianka, 2010). In explanations, theme progression can be used to promote coherence, e.g.: "Pollution can also be a result of natural disasters. For example, hurricanes often lead to water contamination from sewage." In such cases, the news (or rheme) of the preceding clause becomes the theme of the next (Martin, 1992;Schleppegrell, 2004). Also central to the information flow is how technical/abstract wordings are introduced, unpacked and re-packed throughout the text and the overall staged structure of the text (Martin, 2013).
Linguistic choices relating to register variables do not correlate as strongly to genre as generic structures, but they should be seen as equally important for the production and comprehension of disciplinary discourse. In Martin's theory, genre works on a higher level of abstraction than register and coordinates other linguistic resources to achieve communicative goals (Martin, 1992).

Putting Genre Pedagogy into Practice: the Teaching/Learning Cycle
Genre-based teaching is commonly based on a process called the teaching/learning cycle (TLC) (Rothery, 1996;Callaghan & Rothery, 1988). In the version of the TLC which has gained traction in Sweden, instruction is organized in four phases: building field knowledge, deconstruction and modelling of target genre exemplars, joint construction of a target genre exemplar and individual construction of text. While the first phase constitutes an initial and open-ended exploration of the knowledge field, the phases of deconstruction and joint construction are characterized by the identification and joint application of such linguistic features as discussed in the previous section. The TLC ensures a high degree of scaffolding before the students are asked to construct an individual text. It is also an organizing principle used by the teachers in the studies discussed below.

Method
The discussion of different implementations of genre pedagogy will draw upon two empirical studies of genre-based teaching: The first study is a licentiate thesis by Sellgren (Sellgren, 2011). It is based on action research and explores the author's own genre-based teaching in Grade 6 during a curriculum area about factorial explanations in Geography. The findings will be juxtaposed with those of my own PhD thesis (Walldén, 2019a). The materials relevant to the present article were gathered through observations and voice recordings during a curriculum area about maps and population in Grade 6 which lasted for seven weeks. The participant teacher employed genre-based pedagogy, integrating the subjects Geography and Swedish as a Second Language. The empirical findings are analyzed extensively in the thesis and in another article accepted for publication (Walldén, 2009b). Thus, in the present article I will restrict myself to discuss linguistic features of the texts used in this curriculum area. Since both of the mentioned studies focus on genre-based teaching of second language learners in Grade 6, they make for an interesting comparison. The analysis of the texts will draw upon the systemic-functional theoretical constructs introduced in the previous section. The texts cited have been translated from Swedish to English by the author of this article.

Findings and Discussion
Below, an excerpt of a textbook explanation (Haraldsson, Karlsson & Molin, 2008) used in Sellgren's study is shown. It was used for learning about pollution in the relevant curriculum area, and also constituted a model in the deconstruction phase (Sellgren, 2011).
The Baltic Sea currently is one of the world's most polluted seas. Fertilizers from agriculture, exhaust gases from traffic and a lack of sewage treatment works are some of the causes. Since the Baltic Sea is an inland sea, it also takes a long time before the water is exchanged. This makes the pollution which is released there remain for longer. (p 37) Conforming to the generic structure of factorial explanations, the text introduces the phenomenon (pollution of the Baltic Sea) and explains relevant factors. Thus, the generic structure is used for introduction the concept of pollution and unpacking it in more concrete terms. The text also includes technical language (e.g. "pollution", "exhaust gases", "inland sea") and some instances of logical metaphors: "some of the causes", "This makes …". The text construes an authoritative expert voice and seems a suitable model for students' writing in Grade 6. Next, an excerpt of a jointly constructed text is shown (Sellgren, 2011).
The Baltic Sea is a threatened sea and one of the world's most polluted seas. It is a dirty sea because of us humans. One of the major causes of the problem is industries releasing harmful substances which destroy the environment, e.g. carbon dioxide and toxic substances which go straight into the sea through streams and rivers. (p 46) Some instances of repetition (e.g. "sea"), everyday vocabulary (e.g. "dirty"), personal pronouns ("us") and repeated hypotaxis ("which … which …") causes this jointly constructed text to appear less planned, technical and authoritative than the textbook explanation. However, this is to be expected as the current text emerged as a product of teacher-directed whole-class interaction and was likely intended to more closely mirror the kind of writing attainable by the students. It still includes technical language ("threatened sea", "carbon dioxide") and instances of logical metaphors ("causes") and uses the generic structure in a similar manner to the textbook version. As a jointly constructed model before the students' individual construction of text, it has clear merits. Most importantly, the generic structure and other linguistic features of factorial explanations are used productively to engage with content knowledge.
The findings of my own thesis give a contrasting perspective on genre-based interventions. At the initial phase of building field knowledge, there was a clear focus on technical terms relevant to the field of geography such as "climate zones", "precipitation" and "terrains" (Walldén, 2019). There were also abstract terms related to living conditions such as "undernourishment" and "infant mortality rates". However, meanings of these terms and concepts were mediated through spoken language and visual resources rather than texts. This is not unexpected in an initial phase of building field knowledge, and, as the teaching progressed into phases of deconstruction and joint construction, there was an expected shift to written texts. These texts, however, did not draw upon the content knowledge previously negotiated.
The target genre chosen by the teacher was the discussion genre. An excerpt from the first model text she introduced is shown below. In this text, the writer discusses the advantages and disadvantages of moving to a city in northern Sweden. The excerpt shows the two concluding stages of the text: (contrasting) side and position statement.
On the other hand, it is dark and cold for a big part of the year. It is usually between 20 and 30 degrees below zero in January and February. Before the winter really gets going and for big parts of spring, it is slippery for both cars and pedestrians. It is far up in Sweden, so there are not many friends who will want to come for a visit. I think it will be difficult for me to get used to cold and dark Luleå. If I move, it will depend on how good a job offer I get. A good job could offer secur[ity] and make it easier for me to appreciate Luleå.
The text, consisting mostly of everyday vocabulary, offers little in the way of technicality and abstraction. In addition, the subjective orientation of the text ("I think", "easier for me") is less conducive for modelling an expert voice. Even if the text moves through the expected stages of discussions, it does not seem to serve the purpose of advancing an argument in order to convince a recipient. Rather, the writer appears preoccupied with a personal choice.
Apart from modelling a potentially useful generic structure, the text also employs textual themes to guide the reader through the stages: "on the one hand …", "on the other hand …". These logical connections are also prioritised when the teacher leads a deconstruction of the text. Textual themes are valuable resources for organizing discourse in a planned manner according to what is required in written modes of communication, but in this case, they are not used to advance an argument or to engage with content knowledge.
A similar priority in instruction is evident in a jointly constructed text. The topic is whether mobile phones should be allowed during school breaks. As before, the two final stages are shown.
On the other hand, it is not good to use mobile phones during breaks since you could get pointed out on social media which could cause students to feel bullied. There both advantages and disadvantages with using mobile phones during breaks. I have concluded that I think it is a good thing to use mobile phones during breaks.
The use of the textual theme "on the other hand" re-occurs here, and the final stage which sums up the discussion is also marked clearly ("There are both …"). There is also an instance of grammatical metaphor ("cause") and some abstract wordings ("pointed out on social media"). However, this text has even less relevance to content knowledge in geography. In addition, the position statement is not explicitly based on the previous discussion. Just as in the model text, the claims are grounded in the subjectivity of the writer ("I have concluded"… "I think") with little sign of trying to convince a recipient.
In this curriculum area, the teacher also wanted the students to use "linking words" to engage in "developed reasoning". In this sense, the students were also asked to draw upon features of explanations. On one occasion, the teacher constructed criteria for reasoning together with the students. The wording which illustrated "well-developed" reasoning is shown below.
There is not a lot of food in Ethiopia because it is a poor country. This causes many to die because there are not any medicines.
During a lesson which was not observed, the students had watched a movie about living conditions in Ethiopia. My markings of logical connections reflect the ones made by the teachers herself as she wrote down the students' suggestions on the whiteboard. While the wording certainly uses numerous logical connections, including a logical metaphor ("causes"), it does not seem coherent and hardly reflects content knowledge about living conditions in Ethiopia.
By the end of the curriculum area, the students were asked to choose a country from a limited set and write a discussion about whether they would like to live there. Thus, they were required to infuse knowledge about the discussion genre with content knowledge about living conditions. However, and in contrast to Sellgren's study, the modelling of genre structure and logical connections during the phases of deconstruction and joint construction seemed to come at the expense of that content knowledge.

Conclusion
Based on the above studies, it is possible to give some recommendations for genre-based interventions in the teaching of content knowledge. First, it seems desirable to let explicit attention to features of genres and language emerge naturally from the need to engage with disciplinary discourse. In many cases, it can be preferable to depart from the register variable of field rather than from a certain generic structure. My doctoral study (Walldén, 2019a) showed that the discussion genre, which is often associated with quite advanced instances of disciplinary discourse (Martin & Rose, 2008;Christie & Derewianka, 2010;Coffin, 1997), can be trivialized if genre exemplars are tailored to accentuate the generic structure rather than employed for producing meaningful discourse. The generic structure should, as Martin (2001) himself points out, be seen as a probabilistic in relation to the communicative goal rather than deterministic and it only becomes a useful resource when coupled with other appropriate linguistic features.
The theoretical base of genre pedagogy is highly technical, and teachers who seek to implement genre pedagogy, and similar approaches, cannot be expected to grasp all of its complexities. However, I would argue that rudimentary knowledge about the register variables, and the linguistic features associated with them, is a necessary corrective to the restricting fixation on generic structure which can otherwise arise. Apart from field, the analysis of features relating to tenor was revealing as it divulged why some of the model texts failed to model expert voices and how to advance an argument. While it can be very useful to master certain conventions for structuring texts, such as generic structures and textual themes, these features must be employed with a thorough understanding of the relevant field and what the text is to achieve in relation to the recipient.