1In 1988, in a review of work carried out in the field of ESP, John Swales declared:
The scientific paper is one of the most important manifestations of the role of English as the language of wider international communication and yet we still know insufficient about it to be able to provide a proper handbook for those NNS scientists who wish to further their results and their careers by publishing in English. (1999 : 209)
2Since this date there have been many studies on research writing and more particularly the problems faced by NNS (non-native speaker) scientists writing in English. Some studies have analysed the linguistic and rhetorical characteristics of the research article genre (Gosden 1992, Hyland 1994, Salager-Meyer 1992). Both the introduction and discussion sections of the research article have been described in terms of their schematic move structure (Hopkins & Dudley-Evans 1988, Swales 1990) and this has enabled NNS writers to follow the norms and conventions of their discourse communities. St John (1987) and Shaw (1991) focused more on the writing process rather than the product and analysed strategies used by NNS scientists. Several studies have also described the social construction of science and the negotiation which takes place between author, reviewer and editor (Burrough-Boenisch 2003, Gosden 1995, Latour & Woolgar 1988, Myers 1985). In the French context there have been several studies on the problems faced by French scientists publishing in English. Cooke (1993) described the necessary collaboration between anglophone correctors and French researchers in the re-reading of drafts before they are submitted for publication. Birch (1996) examined the first drafts of French researchers to identify recurrent errors. Crosnier (1994) discussed the linguistic criteria used by referees in their evaluations. There have also been contrastive studies describing different conventions in English and French scientific discourse such as Bachschmidt (1999) and more recently Van Bonn and Swales (2007) who describe rhetorical and linguistic differences in French and English abstracts.
3The writing problems of NNS novice researchers have been studied both in anglophone countries (Belcher 1994, Blakeslee 1997) and on the periphery (Flowerdew 1999, Li 2006). It is of interest to see how these less experienced researchers manage to get published and gradually gain knowledge of their community’s writing norms and conventions. Novice researchers may in fact face difficulties writing both in their mother tongue or another language. This is the case of PhD students writing to be published but also secondary school teachers who may wish to embark on a research career by writing a Master’s dissertation or a thesis. The rhetoric of research writing is often picked up gradually as students do more and more reading in their field but it is also transmitted from supervisor to student, often quite implicitly. In order to help students become more effective and autonomous writers we need first to establish what Lee and Swales (2006) call their level of “disciplinary acculturation” that is to say their awareness of the structure and organisation of a research paper and what is expected of them. Research writing is often the result of many drafts and revision sessions with the supervisor, or peers. Thus to have a sample of a student’s own writing and identify its shortcomings we must return to the first version and the successive drafts.
4The aim of this paper is to compare the successive drafts of one research article (RA) to see how and why the language and the content are modified and how a more mature version of the article emerges. Interviews with the student and supervisor also help us to interpret the different choices made. The case study follows a French doctoral student working with a French supervisor to publish in English. However, some of the difficulties revealed could also apply to novice researchers writing in their first language. It is hoped that identifying students’ problems upstream in the writing process will enable us to give advice to those beginning their research writing.
- 1 A standard article has an IMRAD structure : Introduction, Material and Methods, Results and Discuss (...)
5The student in the study attended an English language writing course for doctoral students at Victor Segalen University Bordeaux 2. Pierre (not his real name) is in the third and final year of his doctorate and works in an immunology department. He does not publish in French and the article in question is his first article in English as first author. For this article he began by writing the Material and Methods section and then the Results section1. It is a common strategy for novice writers to begin with the sections that they feel most familiar with. The introduction and discussion were more problematic and he was helped by his supervisor. The article was also re-read by three other colleagues.
6Pierre’s supervisor learnt to publish by reading in her field and working in the laboratory. She is now quite confident and her articles in English are no longer re-read by an anglophone corrector. This is of course quite rare as the majority of French researchers at our university still have their work re-read by a corrector before it is submitted for publication. The modifications to Pierre’s work were largely made by e-mail exchanges with the supervisor adding, deleting, re-writing or commenting on the original version.
- 2 In what follows V1 refers to the first version written by the student and V2 is the supervisor’s mo (...)
7In this section I will look at the changes made by the supervisor to the student’s first draft to see how what Gosden might describe as a “relatively immature, unpublishable piece of writing” is transformed into a “more mature version, now acceptable to the expert RA readers” (Gosden 1995 : 53)2.
8The first version (V1) was too long and was modified to a more concise version which just focuses on the main result and what is new in the study (V2).
V1 : “Engagement NKG2D induces […] T cell cytotoxicity but is not involved in their recognition of CMV-infected cells.”
V2 : “NKG2D triggering induces […]T cell cytotoxicity”.
9As this study deals with a non-native speaker writing in English the manuscript therefore needed correction on a surface level. Before looking at the rhetorical aspects of the discourse, which may pose problems for both native and non-native writers, I will first briefly sum up the local language elements which were revised.
10In the abstract the student’s statement of objective in the simple present was replaced by the preterite to focus more on what had been done in this study and the specific context to avoid too general a statement ;
V1 : “To better understand […] we study […]”.
V2 : “To better understand […] we examined […]”
11In the Results and Discussion section, a simple present verb was replaced by the preterite, this time to limit the results to this specific study and to avoid generalising.
V1 : “The second group is characterized by […]”.
V2 : “The second group was characterized by […]”.
12The student is this study felt that he had problems organising ideas and linking them and indeed in the following example, the student’s choice of link word indicates a relationship of contrast between the two propositions whereas in the supervisor’s version we can see that it is a causal relationship, the second part of the sentence in fact explaining the first part.
V1 : “These results are not restricted to immunosuppressed patients whereas signal triggered by NKG2D alone leads to the same phenomenon on […] T cells isolated from healthy donors”.
V2 : “These results are not restricted to immunosuppressed patients as the signal triggered by NKG2D alone leads to the same phenomenon on […] T cells isolated from healthy donors”.
13The student’s version of the article also lacks signposts for the reader to follow the progression of ideas from sentence to sentence and the supervisor tends to add links to help the reader follow the argumentation.
V1 : “[…] T lymphocytes kill CMV-infected targets”.
V2 : “Moreover these […] T lymphocytes kill CMV-infected targets”.
14When modifications are made to the lexis, they may be words related to academic writing and the degree of certainty with which assertions are made or changes to add more precision to the text. There are also a few cases of L1 interference.
V1 : “We have previously demonstrated that […]”
V2 : “We have previously reported that […]”
V1 : “the population”
V2 : “the worldwide human population”.
V1 : “sensible to”
V2 : “susceptible to”
15The errors made here do not change the meaning but might irritate the potential reviewer or editor.
V1 : “to participate to”
V1 : “[…] relies in the fact that”
16In this section we will focus on changes made to the rhetorical aspect of the discourse. This analysis may thus apply to novice researchers writing in their own language but lacking the experience to know how to organise their research and influence the audience.
17The supervisor in this study said that students had a tendency to focus too much on their own study and did not take into account what had gone before and what was to follow. She underlines the need to specify sources and justify why the work was undertaken. Allison et al. (1998) advocate the use of their Diagnostic Assessment Profile to enable students to analyse their own writing. A key element of their profile for students to reflect upon is ‘substantiation’. Matsuda and Tardy (2007) also stress how knowledge of a topic is based on the representation of others’ work, “one way in which readers assess an author’s rhetorical framing or his or her disciplinary knowledge base is through the use of citations” (2007 : 244). Students tend to be more descriptive and not critical enough and need to “adopt a somewhat more critical stance toward the literature using it to set the groundwork for further inquiry” (Miguel & Nelson 2007 : 81). In this case, in the student’s introduction, references to the work of past studies were at times presented directly as general truths and not as facts which have been demonstrated in the literature.
V1 : “adoptive transfer of […] cells efficiently restricts CMV infection”
V2 : “adoptive transfer of […] cells has been shown to efficiently restrict CMV infection (ref)”
18The supervisor added more specific references to give substantiation to the ideas reported.
V1 : this interaction is independent from […]”
V2 : “this interaction was found to be independent from […] (ref)”
V1 : “Numerous cell clones […] present strong reactivity against CMV infected cells”
V2 : “numerous cell clones have been shown to present strong reactivity against CMV-infected cells (ref)
19The student also had difficulty integrating his own past results. His sentence was wordy and needed an introductory reporting phrase to frame the important piece of information which followed.
V1 : “The marked and persistent increase of circulatory T cells we reported in kidney transplanted patients with active CMV infection strongly pledge for […]”
V2 : “We have previously reported a marked and persistent expansion of circulating […] T cells in the blood of kidney transplanted patients with active CMV infection. This phenomenon strongly supports […]”.
20The student in this study may be influenced by cultural differences and the way things are presented in English as opposed to French. In their study of French and English abstracts Van Bonn and Swales (2007) compared the “deductive, data- and fact-based nature of French prose and the inductive, argument-based nature of English prose”(2007 : 99). They also found that “the French abstracts, however, much less frequently make an effort to justify their research and do not nearly as often reference their work”. (Van Bonn and Swales 2007 : 97). Van Bonn and Swales suggest that French researchers may be used to operating within a smaller academic community and may be more accustomed to giving data rather than arguing a point.
21The supervisor added the sentences below to underline the need for the study and to highlight its importance. As Tardy points out “knowledge-transforming requires writers to engage in the rhetorical act of persuading readers of their work’s volume, significance and credibility” (2005 : 325).
V2 : “There is growing evidence that […] T cells express NK receptors that play a key regulatory role on the effector functions of these cells. Among these, NKG2D is of particular interest in the context of CMV infection”.
22The supervisor also inserted the following sentence to justify the study by showing what remained unknown :
V2 : “The nature of their ligand remains elusive”.
23In a third and final version of the article, the supervisor improved the language, notably by using expressions and lexical chunks commonly found in scientific writing. Certain sentences were added to again emphasise the importance and interest of the study. The phraeseology of these rhetorical moves is much easier for the experienced writer to manipulate. The series of drafts and revisions helps the novice writer gain awareness of the norms and conventions of his disciplinary writing and enables him to pick up expressions which can be used for a future article.
24The abstract thus ends with sentences which underline the study’s importance, centrality and possible applications :
V3 : “These results extend knowledge of...”
V3 : “They also open up new prospects for...”
25In the introduction the supervisor’s aim again is to retain the reader’s attention.
V3 : the ability of... is of obvious interest...”
V3 : “Altogether our previous results pave the road toward the development of new immunotherapeutic approaches...”
V3 : “these analyses could provide new information on...”
26The purpose of the study is thus reinforced and stated more clearly in the supervisor’s version :
V1 : “Hence we analysed the outcomes of NKG2D ligation on […]”
V2 : “Hence in order to assess the functionality of the NKG2D receptor[…] we analysed […]”.
27In the first version of the Results and Discussion section, the student reported his results but made very little attempt to interpret or comment upon these results. There were very few evaluative statements and claims. It is perhaps for this reason that this section was not merely modified by the supervisor but largely added to and re-written. The presence of the author becomes apparent with more persuasive and stronger statements. The supervisor explained that she felt that students tended to stick to their results and did not take a step back to compare with the literature. At the same time the student recognised that his supervisor helped him to make stronger claims with more confidence and to reduce the degree of hedging or cautious statements in the article. Woodward-Kron (2004) expresses this link between caution and persuasiveness ; “Professional academic writing involves mobilising a range of linguistic resources to negotiate with prior texts and persuade the community to accomodate new and possibly conflicting claims” (2004 : 141). We thus find that the supervisor adds expressions which give the author’s attitude towards the findings :
V2 : “as expected […]”
V2 : “by contrast […]”.
28The supervisor also adds expressions in order to compare the results with those of the literature :
V2 : “this result was consistent with […]”.
29There are tentative explanations and admission of unexpected outcomes :
V2 : “their resistance […] may be related to […]”
V2 : “surprisingly no difference could be observed”
V2 : “these results suggest that […]”
V2 : “thus, there is a very close similarity between results obtained […] and those obtained […] demonstrating a clear, direct, activating function of […]”
V2 : “such […] has previously been reported only for […]”
30In the third version of the article, the supervisor enhanced the Results and Discussion section again by attempting to explain certain factors and by indicating potential applications of the study :
V3 : “... is an attractive hypothesis”
V3 : “... could also be envisaged as an alternative way to...”
31In his investigation of a corpus of student writing, Hyland (2004) found a relative absence of lexical bundles expressing stance and engagement. His explanation for this is that student writers “may be uncomfortable in explicitly aligning themselves with a particular evaluation or personally attesting to the weight they want to atrribute to their claims” (Hyland 2004 : 16). The supervisor in his more mature version of the text is able to take a stance, evaluate the material and engage with the reader more directly.
32In the introduction more information and references were added by the supervisor to give more context, background and to better situate the author’s own study.
V2 : “Accordingly a delayed expansion of […] T cells in the course of CMV infection is associated with increased and longer viremia and with more symptomatic infections”.
V2 : “transcript expression of these proteins is increased in CMV infected cells […]” (refs).
33In the Material and Methods section, a paragraph was added while another was removed.
34The first paragraph of the Results section was removed so that the section then began directly with a subtitle which summarises the main result.
V2 : “NKG2D engagement on […] cell clones does not induce […]”
35A paragraph was added to review what was known from the literature, to give a reminder of the author’s own past findings, to underline again what they had found in this specific study.
36A comparison of the author’s final version and the published version also reveals that the editor further improved the language by re-writing in a more formal academic style :
V3 : “pave the road toward...”
V4 : “pave the way toward....
V3 : “...is a pre-requisite to be gone through”
V4 : “...is a necessary pre-requisite”
V3 : “We wondered whether.....
V4 : “We sought to determine...”
V3 : “This allowed to reveal the effectiveness of...”
V4 : “The effectiveness of... was thus revealed.”
37Hyland’s study of frequently occurring word sequences or lexical bundles in a corpus of scientific writing revealed that “it is often a failure to use native-like formulaic sequences which identifies students as outsiders and there is a general consensus that formulaic sequences are difficult for L2 learners to acquire” (2007 : 4).
38Through this case study it is apparent that the majority of the changes made by the supervisor were in the Introduction and Results and Discussion sections and concerned especially the management of the discourse. It would seem that in the introduction to a research article, the most important aspect for the newcomer is to know how to refer to the literature and situate their own work within what has gone before, to be able to underline their work’s significance and to clearly state their objective. In the Results and Discussion section the novice writer should take a stance to interpret and comment upon the findings. The results here are consistent with Gosden’s study (1995) of novice researchers in Japan where he cites deletion and addition of technical data, reshuffling of statements and ‘rhetorical machining’ including the structure of the discourse and the language of claims and purpose as the main modifications carried out. Belcher reminds the reader that reviewers are looking for topic relevance, the novelty of the contribution, subject matter familiarity, research design and data analysis and presentation and style (2007 : 7). It thus seems logical that in his review Flowerdew (1999) states that successful writers are able to measure their research against current conversations in the discipline, relate to the anticipated audience, structure their argument appropriately and use appropriate degrees of persuasive language.
39The supervisor’s guidance and the student’s imitation of her more mature style of writing led to the paper being accepted for publication without further revision. Their collaboration is of interest as it shows how, in the French context, a non-native speaker student and supervisor work together to publish successfully in English. What can be learnt here on the rhetorical framing of research can also be used by novice researchers writing in their first or other language to enable them to imitate the more mature style of experienced researchers.