Citation practices of postgraduate students writing literature reviews

Writing a literature review requires highly sophisticated academic literacies. Many postgraduate students find this genre a challenge. While there is a growing awareness of the need for explicit pedagogy to support students writing this genre, many pedagogical interventions fail to move beyond a focus on citations as a stylistic convention or as a way of avoiding plagiarism. What is missing is a pedagogy that relates citing to the more complex, fluid conceptual and ontological practices that are implicit in academic contexts. The purpose of this paper is to explore the citation patterns, complexity and discursive practices in master’s students’ literature reviews, and to inform pedagogy.


Introduction
One of the most complicated writing tasks for postgraduate students is the literature review. Many postgraduate students experience extreme anxiety when writing these papers or chapters (Green and Browser, 2006). Even experienced writers struggle with this cognitively complex genre (Webster and Watson, 2002). The literature review involves selecting sources, reading critically, extracting from texts, and synthesizing sources into writing through citations. All these tasks require an understanding of the multiple literacy practices of academic discourses. While there is an awareness of the need for postgraduate students to master these literacies, and a body of research that provides a growing empirical base, many pedagogical interventions fail to move beyond a focus on citations as a stylistic convention or as a way of avoiding plagiarism. Missing from this is a pedagogy that relates citing to the more complex and fluid discursive practices that are implicit and deeply embedded in academic contexts. The purpose of this research was to explore the discursive nature of citation practices and patterns in master's students' literature reviews, to understand the literacies involved better, and to inform pedagogy. This paper makes the argument that a pedagogy incorporating the conceptual and ontological nature of this genre would enable postgraduate students to engage with the literature review from the position of insiders, rather than onlookers from the periphery.

Literature reviews
The literature review is a demanding genre to write. Completing a literature review successfully requires a set of significant intellectual tasks. Literature reviews involve a level of mastery over existing published research in a discipline or field. Writers must articulate contradictions, gaps, inconsistencies, and the relationship to the problem under examination (Vardi, 2012). They must also make decisions about sources in terms of breadth, selection, relevance, currency, availability and authority. Literature reviews demand sophisticated intellectual literacies and high-level thinking skills (Bruce, 2001;Turner and Bitchener, 2008). In addition, writing a literature review is a complex process of constructing language by using others' voices to establish kinship and authority within a discourse community (Thompson, 2005). Literature reviews are becoming a central focus in postgraduate student education, and calls for more attention to be paid to pedagogical issues around this genre are increasing (Kamler and Thomson, 2006;Randolph, 2009;Turner and Bitchener, 2008).
In a recent analysis of the literature on literature reviews, Chen et al. (2016) grouped current research along four lines of focus: Under the linguistic category, Chen et al. (2016) assemble the research that identifies the problems students experience as primarily a challenge of language: vocabulary, sentence patterns, genre requirements, rhetorical moves, citation patterns and so on (Feak and Swales, 2009;Kwan, 2006;Turner and Bitchener, 2008).
Methodological challenges refer to the difficulties students have in selecting relevant papers for review, extracting necessary information they need and synthesizing it all together in a coherent argument (Granello, 2001;Holbrook, 2007). In this category, the emphasis is on methodology: what strategies do students use when they are unsure of what to read, how widely to read, what to choose and what to take from the readings chosen? Chen et al.'s (2016) third category shifts the focus to conceptual challenges. Conceptual challenges relate to how students understand the purpose of the literature review, such as identifying gaps in the literature, providing a rationale for the study and evaluating the literature in the field in order to situate the study. Research in this category found that students tended to focus on lower-order thinking skills, such as finding a list of any papers related to the topic rather than globally scanning the literature and then deliberately selecting. Significantly, many new researchers were not aware of the reflexive nature of the literature review. By this, Chen et al. mean the iterative nature of reading, rereading, reflecting, adding and building an understanding of the literature -a process that is usually not linear. This group of researchers stressed the tangled and ambiguous path that writers often follow when the initial identification of the literature proves to be superfluous after the first readings and further literature needs to be sought. In addition, this research emphasized the way that, in developing a literature review, knowledge is constructed through reflection. Building blocks for the rest of the paper are laid, torn up, rebuilt and developed as the writer grapples with understanding the literature. Ultimately, the literature review encompasses different phases of thinking that impact on the rest of the paper (Kwan, 2008). Authors who discuss conceptual challenges include Bitchener and Basturkmen (2006), Boote and Beile (2005), Holbrook (2007), Kwan (2008Kwan ( , 2009, and Suri and Clarke (2009).
The final category identified by Chen et al. (2016) in their analysis of literature reviews concerns ontological challenges. Challenges of an ontological nature are essentially about identity, knowledge construction and participation in discursive practices. The focus of research shifts here from language, rhetoric and genre to understanding how writing a literature review involves identity-work by engaging in the epistemologies of the discipline through the writing process itself. For example, confidence and authority in writing are developed through a growing knowledge of the literature, and enacting that through continued writing practice. While failings are often attributed to poor writing skills or language, weak methodological knowledge or a lack of understanding about the conceptual purpose of the literature review, newcomers can hardly be expected to feel competent critiquing the literature when they are positioned as outsiders to the discourse. Writing a literature review involves performing an identity as a scholarly author, which is developed through sustained engagement with discourse communities (Hyland, 2008). Understanding the discursive nature of literature reviews -how much it positions a writer in a discourse -is an aspect of this genre that is often not explicitly explained. Novice writers are left to discover this through their own hit and miss experiences. Authors that discuss ontological challenges include Golde (2007), Kamler and Thomson (2006), Ruttan (2004), Kwan (2009), Qian andKrugley-Smolska (2008), and Warburton and Macauley (2014). Chen et al. (2016) show that recent research on literature reviews is weighted heavily in favour of the linguistic and methodological challenges that students face, while fewer studies focus on the conceptual and ontological challenges. Framed by an academic literacies approach, this study aims to address this gap, particularly in relation to how students are taught to write the literature review.
Academic literacies (Lea, 2004;Street, 2006, 2014) seek to explore student writing from a perspective that critically engages the social and institutional contexts within which writing is located, rather than automatically locating the 'problem' within individual deficit (Baker, 2017). This approach seeks to understand the range of literacies required of students to succeed at university (see, for example, Blommaert and Horner, 2017;Clarence and McKenna, 2017;Kaufhold, 2017;Lillis et al., 2015). Using the term 'literacies' instead of 'literacy' emphasizes 'contested sets of socially situated writing practices', which develop historically over time and are located in space (Lea, 2016: 89). The literacies required are heterogeneous, diverse, sometimes tacit and often not obvious to students (Kaufhold, 2017). Academic literacies are also an ideological approach, in that literacies are not viewed as superficial communication skills that need to be acquired but rather are deeply held, sometimes conflicting epistemologies that exist in academic disciplines. When students participate in the literacies of their discipline, they begin to develop their identities in that discipline. An academic literacies approach suggests that disciplinary literacies should be explicitly taught in university contexts, including contestation over what these literacies are. This pedagogy should empower students to be active in their own learning and to be transformative rather than to conform to prevailing norms (Lillis and Scott, 2007).

Research context
The study was conducted among master's student writers in the Faculty of Education at a medium-sized university in Canada. Master's students in the postgraduate programme have the option to choose from four routes -thesis, internship, portfolio or all-coursework -in four different domains of study -teaching and learning, educational leadership, information technology and post-secondary studies. Most postgraduate students in the programme work full-time and have demanding careers as educators. Consequently, the all-coursework route is a popular choice for many students enrolled in the programme. A culminating course -the Research and Development Seminar -is the final requirement in the all-coursework option. In this seminar, students are expected 'to explore in greater depth a topic of interest and personal relevance' (Hammett and Collins, 2002: 440). This final course is designed around supporting students to write a substantial literature review paper on a topic of their choice. Students are required to develop a proposal, a draft and a final paper, although they could prepare more than one draft of the paper if they wished. The proposal and draft(s), once written, are subjected to 'critical dialogue' from groups of three or four peers. Critical dialogue refers to 'questioning the basis for assumptions and assertions, critiquing the research question, examining the construction of argument, or noting appropriate application of American Psychological Association publishing conventions' (ibid.: 441). This critical dialogue was set up to mimic the process of peer-review of articles submitted to scholarly journals.
These students were taught methodological and some linguistic aspects of literature reviews, including how to search and select source papers and how to establish a research question. Although they were told to develop an argument and be cognizant of counter-arguments, they were not explicitly taught how to construct an argument. There was a strong focus on correct citation conventions, and avoiding plagiarism was taught as a key priority.

Methodology
Data were collected in the form of student papers (drafts and final versions) from this final seminar course, with the aim of analysing the literature reviews through citation patterns and practices. From a total of 179 students who were invited to participate in the project, 50 (28 per cent) agreed and 23 (13 per cent) submitted both draft and final papers. Data analysis of the 23 draft and 23 final papers involved: 1. Tracing journal citations in each of the papers to check the reliability and use of sources.
Journal articles were the most often used sources and book sources were only tracked if substantial use was made of the source. 2. Analysing each citation (number of references per paper; how references were used in the paper). Each citation was coded for reliability of citation (accuracy, attribution, traceability); use of integral (author prominent) non-integral (research prominent) citations; single study (one source) versus group citations (many sources grouped together); use of direct quotations (quotations of less than 40 words used in a paragraph) or block quotations (indented quotations of more than 40 words); use of global summary (summary of entire paper) or paraphrasing of specific parts of the original text (references to sentences or passages); and presence of plagiarism. All these citation practices were drawn from Feak and Swales (2009), and are considered required academic citation practices for graduate students. In each paper, every citation was traced and coded according to the above categories. For example, 'paraphrasing' refers to the citation that followed the paraphrased sentence. 3. An analysis of higher-order thinking skills using Bloom's taxonomy using Granello (2001) to establish levels of cognitive complexity in the literature reviews. Using the older taxonomy (knowledge, comprehension, application, analysis, synthesis and evaluation), we used Granello's descriptors of how these appear in graduate student literature reviews to code complexity. 4. The papers were also analysed using the constant comparison method (Merriam, 2009;Savin-Baden and Major, 2013). This process involved: (1) Reading the papers -the papers were read/analysed iteratively until the data were saturated; (2) Coding and thematic analysis -in each reading of the papers, citations were coded (for example, integral or non-integral citations), presence or absence of higher-order thinking skills were coded (for example, synthesis present), and keywords or phrases noting contextual information were recorded, (for example, flexible citation use); (3) Connecting strategies -codes were grouped into tables and keywords/phrases into themes, and compared for similarities and differences. The papers were also compared to each other and across the data set to establish the validity of themes and to ascertain universal patterns and commonalities. The value of the constant comparison method is the continual comparison of coded and uncoded data to find consistencies and differences. This process allows for the meaning within each paper to be retained, while at the same time examining patterns across the papers.
Analysis of the data centred on addressing two questions: • What citation patterns, complexity and discursive practices were evident in these literature review papers? • What could these patterns tell us about the literacies that students engaged in?

Continuum of citation use
The result of the process of analysis described above was that participant writers were grouped along a continuum from high to low intertextual flexibility. The concept of intertextuality here draws on the postmodern notion that each new text is constructed from other texts, which have a history and are traceable. In essence, it refers to how writers use previous texts to create a new text and is the representation of others' writing in one's own writing (Prior, 2004). Nine student writers were grouped towards the high end, another nine towards the low end and five ranged in between.

High intertextual flexibility
Students towards the high end of the continuum demonstrated higher-order thinking skills (analysis, synthesis and evaluation) when citing sources in their papers. They usually had a holistic understanding of each source and provided global summaries of a source paper or author's perspective. They tended to summarize or paraphrase the original texts in their own words instead of keeping the same words or phrases as the original texts, or practising patchwriting (Pecorari, 2008). Citations in these papers tended to reflect students' own interpretation of sources, and they knew how to make use of sources to support their arguments. These writers used direct quotations and block quotations only when necessary, and the quotations did not constitute a large percentage of the total number of citations. They tended to collect together sources with similar ideas and used more group citations in their papers than others elsewhere along the continuum. Not all writers exhibited all the characteristics described here, but overall they showed a flexibility in their use of sources. For example, they used citations to persuade and convince, to align with particular perspectives and to connect to particular cultures, and they were able to repurpose source texts to fit a new line of argument (Hyland, 1999(Hyland, , 2004(Hyland, , 2008.

Low intertextual flexibility
These writers demonstrated fewer higher-order thinking skills when citing sources in their literature reviews. They usually tended to cite information from specific sentences in source texts and often it was clear that they did not have a holistic understanding of each source. They kept close to the words of the original texts in their own citations, in some cases so close that they could be considered plagiarized. Direct quotations and block quotations made up a large percentage of the total number of citations. These writers rarely used group citations; instead, they cited each source separately, usually leading from the author (integral citations). Some of them tended to mine information from the abstract of the source paper or the first few paragraphs of the article. There were often misrepresentations of the original sources and many citation errors. These writers were less flexible in their use and arrangement of sources.
Although the student writing ranged along a continuum, the two ends suggest a binary between 'good' students (high intertextual flexibility) and 'weak' students (low intertextual flexibility). This labelling and the binary hide a deeper complexity, and a closer examination of these practices in the context of each paper yields some interesting results about the shades and tones of this genre and the literacies required, and about the effects of pedagogy. Two cases, Armani and Caelan, will be examined in detail to illustrate these complexities. (Pseudonyms have been used to protect participants' confidentiality and anonymity.) These cases are illustrative of the two groups, but are presented here to showcase the finer brushstrokes that are lost when viewing the bigger picture. Armani was identified as a writer with high intertextual flexibility. She chose to research and write about standardized assessments in the K-12 school system. She reworked the draft paper substantially by adding many more citations and sources, and fleshing out her argument with supporting details (see Table 1). While the draft contained more direct quotations, many of these were changed into paraphrases in the final paper. The bulk of the citations, 62 per cent, were non-integral (research focused), while integral (author focused) made up 38 per cent of the citations (see Table 2). Most of her citations were single-study citations and she only grouped together citations in two cases (see Table 3). There is little evidence of plagiarism and her citations are characterized by global summaries of papers, rather than by paraphrasing parts of source papers or quoting directly (see Table 4). Her final paper only contained three direct quotations out of a total of 90 citations. a Each citation was coded as integral or non-integral. b Where the author is prominent in the citation, for example 'Jones (2017) argues for increased standards'. c Where the author is only named as a reference, for example '… increased standards (Jones, 2017)'.  The pattern for Armani's citations was to summarize articles in her own words and then add her understanding to it. Often students retain certain keywords from the original sources, but Armani tended not to keep keywords and she provided summaries in her own words. She rarely cited from abstracts or from sentences in source articles. Her citations were precise and accurately followed the required referencing style conventions.

Case 1: Armani
From her papers, it is clear to see that Armani had strong views on standardized assessments, no doubt based on her experience of teaching. In the draft, this argument is more direct. In the final paper, the argument has been qualified by including research that showed a more complex picture of standardized assessments, perhaps reflecting the review process by her peers. Also, by qualifying her argument she showed that she had become aware that academic writing often contains counter-arguments. The structure of her paper showed knowledge of the genre and an understanding that literature reviews are conceptual. There is a clear purpose, rationale and guiding question. As a teacher, she is present in the paper through her argument. As a student writer, she is also present in the paper through her understanding of the genre.
Here is an example of one citation (Example 1):  (Starch and Elliot, 1913) and presents studies showing the documented weakness in typical classroom assessment practices (Gullickson and Ellwein, 1985).
This example shows that Armani moved from a bold, almost truth-like statement based on one citation to a more detailed paragraph that extends the citation substantially. The citation shifts from non-integral to integral/author prominent, which indicates a conceptual shift from 'truth' to argument. When author-prominent citations are used, there is more credence given to ownership of ideas, as opposed to the truth-like facts of non-integral citations (Hyland, 1999: 355). In the final version, the cited author is clearly positioned as making the argument: he 'states', 'argues', 'counters', 'points out' and 'supports'. The author and citation is present and at the forefront. Armani makes an effort to build the argument that Cizek makes by adding details: 'more reliable, more relevant, less biased'. She makes the point that Cizek supports his claim and mentions the sources he draws on. The change from draft to final paper also involves moving from a direct quotation to a larger paraphrase. Essentially, this is a shift from information-seeking to argument-presenting. While the language of the quotation -'the contention that the quality of teacher-made tests pales compared with the more rigorously developed, large scale counterparts' -is still present in the paraphrase, and may possibly be seen as plagiarism, the expanded paragraph reflects more understanding of the source material. In Example 1, Armani has drawn from three different parts of the original text (1, 2 and 3 below) and rearranged them to suit her purposes: Original source: 1. It is not an overstatement to say that, at least on the grounds of technical quality, the typical highstakes, state-mandated test that a student takes will -by far -be the best assessment that student will see all year (p.25).

On the one hand, there is evidence that concerns about the extent (Phelps, 1997) and cost (Phelps, 2000) of testing are overblown, and there is evidence that the depth of anti-testing sentiment in the populace has been overstated (Business Roundtable, 2001) (p.19).
While the words 'decades of evidence' and 'rigorously developed large-scale counterparts' are from the original, it is clear from Armani's writing that these are Cizek's words. The paragraph overall reflects a global understanding of the source. This pattern of one sentence or claim in the draft being expanded into a paragraph with more detail and support from the source article in the final paper is characteristic of Armani's writing. In the final paper, she also added 18 further sources to expand her argument and to include the counter-argument.
Here is another example (Example 2): Armani, draft: In general, the public's attitude toward testing is favourable yet chronic opposition persists when testing is attached to consequences such as student placement, promotion, retention, and bonuses for teachers (Cizek, 2001).
In the final paper, Armani added content and citations to produce this: Armani, final paper: In general, the public's attitude toward testing is favourable yet chronic opposition persists when testing is attached to consequences such as student placement, promotion, retention, and bonuses for teachers (Cizek, 2001 de-skilling (Volante and Cherubini, 2007); and in narrowing or simplifying of curriculum (Wang et al., 2006).
Armani is performing a scholarly identity here. Several citations are grouped together and refer to a number of issues. This shows synthesis of readings and knowledge of scholarly ways of citing. Examining the original sources provides a glimpse into how deeply Armani read and understood the source articles. 'Higgins, 2009', for example, does not explicitly address socio-economic status. However, it is implied throughout the article and is an implicit underlying theme. 'Volante and Cherubini, 2007' do mention deskilling and teacher inexperience, but these are really smaller points in the article, which focuses on leadership. In 'Wang, et al. 2006', the original text states: 'They further argue that standardized tests oversimplify knowledge and do not test higher order thinking skills' (313, emphasis added). The article sets out many of the issues and debates, and the key criticism in the article is about the 'one size fits all' approach of standardized tests. This is a correct interpretation of curriculum narrowing by Armani. The point here is that Armani has moved away from the language of the source texts and is making global interpretations of meaning. She has repurposed and recontextualized source texts to support her argumentevidence of her flexible use of citations. Overall, the final paper is a culmination of Armani's understanding of the conceptual nature of literature reviews and referencing conventions, with her teaching experience and passionate views on standardized testing that have enabled her to conduct deep reading of the articles. She has shifted from her position of one-sided truth-like statements in the draft to including counterarguments and shifting to argument-like statements in the final paper. She writes as a knower, not from her position as a postgraduate student but from her position as an experienced teacher: she knows her subject. Armani has high intertextual flexibility because she exhibits agency in her citations. She uses citations to illustrate, to provide evidence, to counterargue. In the final paper (unlike in the draft), she manoeuvres quite successfully through the sources, linking citations with suppleness and agility, and exhibits the ontological nature of literature reviews. Case 2, Caelan, shows a different engagement with writing a literature review. Caelan was identified as a writer with low intertextual flexibility. The focus of her paper was differentiated instruction in the K-12 schooling system. Like Armani, Caelan's paper changed substantially from draft to final stages (see Table 5). She restructured much of the paper from draft to final paper. Sections were moved around, the paper was reorganized and new headings were created. She more than doubled her sources in the final paper. Caelan also had strong views about her chosen topic, no doubt related to her experience as a teacher. a Each citation was coded as integral or non-integral. b Where the author is prominent in the citation, for example 'Jones (2017) argues for increased standards'. c Where the author is only named as a reference, for example '… increased standards (Jones, 2017)'. However, her final literature review, while containing a strong argument, is more tentative. This tentativeness is performed through citation use. Caelan tended to use non-integral citations (see Table 6), which indicates that she did not foreground the positioning of authors or show explicit arguments. She used single citations and did not group citations at all (see Table 7). Each citation was used as a separate entry. Final paper F2012-7F 41 (89%) 2 (4%) 0 1 (2%) 0 a Citations were coded for direct quotations, block quotations, global summary, paraphrases and plagiarized citations. The figures here are percentages of the total citations. b Quotations of less than 40 words, placed in the text. c Indented quotations of more than 40 words. d Citations that follow a sentence/paragraph summarizing the entire source paper. e Citations that paraphrase a sentence/paragraph in the original source. f Citations that follow a sentence/paragraph that is identical to the source.

Case 2: Caelan
A key characteristic of Caelan's citations is that she used direct quotations for the majority of her citations (see Table 8): 41 of the 46 citations are direct quotations, despite the explicit instructions from the course facilitator not to rely too heavily on direct quotations, and the peer review process, which was also mandated to curb direct quotations. Caelan's writing reflects a writer who is acutely conscious of referencing conventions and not breaking the rules, and the possible impact of assessment. Consequently, she yields her citations with less agency, little risk and an almost tangible fear of moving too far from original sources.

Caelan, draft: Differentiated Instruction is a responsive instructional method, as it allows teachers
to become more increasingly proficient in understanding their students as individuals, increasingly comfortable with the meaning and structure of the discipline they teach, and increasingly flexible in order to match instruction to students needs with the goal of maximising the potential of each learner in a given area (Tomlinson, 2003). 'Teachers may adapt one or more of the curricular elements (content, process, product) based on one or more of the student characteristics (readiness, interests, learning profile) at any point in a lesson or unit' (Tomlinson, 1999, p.11). 'Differentiation constitutes an innovating, constant, reflective procedure of effective teaching and learning that cannot be met by readymade lesson plans' (p.3).
Caelan, final paper: Differentiated Instruction (Tomlinson, 2003) is a responsive instructional method as it allows teachers to become more proficient in understanding their students as individuals, more comfortable with the meaning and structure of the discipline they teach, and more flexible in order to match instruction to student's needs. The goal of the teacher is to maximise the potential of every learner. Differentiated instruction develops in different ways … Contrary to the one size fits all approach to lesson planning, 'differentiation constitutes an innovating, constant, reflective procedure of effective teaching and learning that cannot be met by readymade lesson plans' (Stavroula, Leonidas, and Mary, 2011, p.3).

Tomlinson's (2003) original text: Differentiated instruction is responsive instruction. It occurs as teachers become increasingly proficient in understanding their students as individuals, increasingly
comfortable with the meaning and structure of the disciplines they teach, and increasingly expert at teaching flexibly to match instruction with student need (p.3).
In the draft, Caelan collects together quotations and close paraphrasing from the Tomlinson citation to construct her paragraph. There is little meta-discourse or author attribution to situate the argument as Tomlinson's. Statements carry a 'truth-like' tone. The final version of this same paragraph shows her attempts to build on the draft paragraph. I have excluded a long block quotation in the middle, which is now represented by an ellipsis. What is most notable is that she has moved the citation right up to the beginning of the paragraph, where it is tacked on to 'Differentiated Instruction' rather awkwardly. One can see how she has tried to ensure correct attribution, possibly to counter any accusations of incorrect citing and plagiarism from her peers or assessor. Her paraphrasing of Tomlinson's original text contains only minor changes of words. The final paragraph changes little except for additional comments, two new sources and some cutting.
When Caelan did paraphrase, to ensure she did not plagiarize, she changed her version to the point where her words often did not accurately reflect the original source: Caelan, final paper: Adams and Pierce (2004) suggest that there are four elements of DI instruction. They include: classroom management, anchoring activities, DI activities and assessments.
Original source: Essential elements for successful differentiation include specific classroom management techniques addressing the special needs of a differentiated classroom, planned use of anchoring activities, and flexible use of time, space, and student groups (p.1).

The significance of pedagogy
This study sought to examine the citation patterns, complexity and discursive practices evident in these literature review papers, and what these practices could tell us about the literacies that students engaged in. For participants, completing the all-coursework master's degree option meant that, unlike thesis students, these students rarely saw themselves as producers of knowledge, nor did they feel that they were entrenched members of an academic community (Hammett and Collins, 2002). Both Armani and Caelan received the same instruction, and both were confident about their topics. Yet in the process of moving from draft to final paper, Armani was able to shift from truth-like statements to positioning an argument through her citation use. Caelan, on the other hand, became stuck in the mire of avoiding plagiarism. In effect, Caelan performs the 'good' student because she reflects the existing pedagogy, while Armani is the outlier for understanding (whether consciously or not) the implicit discursive nature of the literature review.
Although these literature reviews are snapshots, they reflect ongoing processes that collect together in time and space. In this moment, the text is an organizer of relationships. What comes together in each of these texts are the identities of each student writer and the discourses they draw on: an experienced teacher in the K-12 system, a master's student in a post-secondary context and an academic writer. Appearing in each paper is a gathering of the research subject and knowledge of it, teacher identity with opinions and experience of practice, the teaching and learning of literature reviews, the rules around plagiarism, the rules around citations, the rules around writing the literature review (genre moves), expectations of assessment, a writer identity, and a scholarly identity.
These findings highlight the varied relations between individual writers, the text and academic discourses. In these relationships, the writer not only acts but also is acted upon. The two cases illustrate how Armani acts on her literature review and how Caelan is acted upon by the impending doom of plagiarism. Armani is able to use citations flexibly to repurpose sources for her argument, to showcase her scholarship, to suit her audience and to perform her identity as a scholarly writer. The literature review is a mechanism to engage in debates around a topic she is passionate about. Caelan, on the other hand, is reluctant to leave the safety of her sources. The rules around citing and the fear of plagiarizing (and the consequences for assessment) loom large in her literature review, despite her knowledge and passion for her topic.
Academic writing is heavily shaped by prevailing discursive and contextual practices. The ability to incorporate voices from other texts into one's own writing is essential to the construction of one's own knowledge and is influenced by what counts as knowledge in a discipline. The conceptual and ontological challenges outlined by Chen et al. (2016) align with an academic literacies approach that advocates that it is difficult to understand the struggles students experience unless we locate writing within social and discursive practices, and ultimately find a way to facilitate the complex gathering of writer, reader, texts and discourse.
Constructing an intertextuality of multiple voices in writing implies that knowledge is not simple or transparent. Rather, gathering multiple sources into a coherent argument indicates a more dialogic approach, an assemblage of voices, positioned in particular ways by the writer. It also implies an epistemology of knowledge as contested and constructed by the writer. In other words, for students to understand how to use citations and sources, it is critical for them to understand how knowledge exists -usually as debateable and constructed -in academic contexts. However, the use of textbooks at high school and sometimes undergraduate levels, particularly in North America, often reinforces the idea of transparent knowledge based on the voice of an uncontested expert (Hendricks and Quinn, 2000). Research shows that students find it immensely challenging to detect authorial stance in their readings. It is not surprising, then, that they find it difficult to differentiate between their views and the views of those they are reading, and to integrate their own voice into other voices (ibid.). For many students, knowledge is 'out there', and merely needs to be reported on.
In addition, assessment practices in the transition from high school to university, and sometimes in undergraduate courses, implicitly fortify the perspective that knowledge is transparent (Baker, 2017). Baker (2017: 15) makes the argument that assessment 'is both consistently a dominant discourse and "the driver" for academic writing'. In other words, students align their understanding of referencing to assessment requirements. What is read, how much is read and how it is cited are shaped by a need to satisfy assessors, rather than as a way of constructing knowledge (Hendricks and Quinn, 2000). While many of the rhetorical, genre and writing processes in literature reviews are being taught more explicitly, many of the discourse practices still remain invisible (Chen et al., 2016). This further underscores Hathaway's (2015) argument that writing pedagogy should not be seen as 'support' for some needy students, but as a necessary entry point for all students and as located in the mainstream of academic culture.
What does this mean for pedagogy? If we want students to write like Armani, we need to explicitly show novice literature review writers the discursive nature of this genrethe ontological challenge in Chen et al.'s (2016) terms. This new pedagogy would include: (1) discussions about notions of 'truth', argument and the different perspectives of the source articles, as well as the writer's own standpoints; (2) an engagement with how writing a literature review involves performing an identity as a scholarly author (Harwood and Petrić, 2012) -as Hyland (2015: 5) has argued in the case of experienced writers, it is through identity that academics 'both achieve credibility as insiders and reputations as individuals', and it is through citations and sources that that we align ourselves with particular groups by taking on their 'discourses, genres and understandings' (ibid.); and (3) how literature reviews involve negotiating a subject position on the page (James, 2014), and that this is a choice of who students want to be, and how they want that subject positioning to appear in the text. In addition, student writers need to know that, as James (2014: A-15) argues, the process of creating a writing subject is a 'fragile and at times contradictory process', and characterized by 'agency within discursive constraint'. It is through a dynamic process of awareness, considered writing, and numerous revisions that a writer's subject positioning emerges.
Pedagogy tends to focus on the linguistic and methodological literacies of the literature review, and on citation conventions that revolve around avoiding plagiarism, instead of critical engagement with source texts, writers and academic discourses (Vardi, 2012). As Lea (2016: 90) suggests, 'working with conventions alone does not engage students in contested knowledge making'. How experienced writers choose to present information through citations is as important as the information they present (Hyland, 2004). An explicit pedagogy would attempt to overcome the separation of the conceptual/ontological and linguistic/methodological activities, while encouraging questions around positioning, agency and identity performance. It is through knowing these literacies -the discursive nature of literature reviews -that student writers can openly negotiate multiple micro-decisions, and enhance their capacity to grasp the elusive nature of literature reviews.