English or Englishes in Global Academia: a Text-historical Take on Genre Analysis

The challenge of publishing internationally for non-native English speakers (NNESs) is substantial, although there are conflicting accounts as to how NNES-authored texts fare in English-medium journals and the nature of the criticism levied at these texts. Collaborators from a wide variety of backgrounds and skill sets may contribute to these texts, and the aspects they focus on differ based on their profile. One of these aspects, rhetorical appropriateness, is of interest to the study of NNES writing because of difficulties authors have in adapting to the discourse-level features of English-medium academic texts. This article presents a multi-year research project exploring the rhetorical characteristics of writing produced by 10 NNES academics seeking to publish in international biomedical journals. Using a text-historical approach, the study traces the arc of 10 different research articles across multiple drafts, analyzing the processes and agents behind these drafts and the feedback received from target journals. Focusing on rhetorically significant changes made across different drafts and comments concerning linguistic issues, this paper seeks to further the understanding of English as a lingua franca within written discourse in the field of biomedicine. One text history is presented to exemplify the methods.


Introduction
In recent decades, the dominance of English as a medium for academic text production and the premium placed on English-language publication have been widely documented [1][2][3].Faced with the veritable need to publish in English, certain studies have found that non-native English speaker (NNES) researchers are at a disadvantage on this English-only playing field [1,2], although other reports gather author perceptions that partially challenge this claim [4][5][6][7].Real or perceived, the advantage enjoyed by researchers who are native English speakers (NESs) is not always considered by NNES authors to be an unfair one [8].While some researchers state that English usage and style are not the reasons given by journals for rejecting submissions [9], others say just the opposite (Weller, 2001, cited in [10,11]).Two of the primary obstacles that make the goal of publishing in Anglophone academic journals such a challenging one for NNES academics are time [12] and cost [2,12,13], both of which are at least partially attributable to linguistic background.
In order to improve the linguistic quality of their texts, NNES authors may seek support from a variety of individuals.Lillis and Curry adopt the blanket term literacy brokers [3](p.4) to refer to the wide range of figures who influence academic texts, from academics from within the same academic field as the aspiring author, to language professionals such as translators, copy editors, proofreaders, and other English-language specialists, and even nonprofessionals such as friends, spouses, or other family.Burrough-Boenisch [14] devotes a section of her article to the numerous ways of referring to English-language specialists who shape academic texts before finally settling on the term "authors'editor".Over 30 years ago, Tacker [15] delineated the roles played by author's editors within multilingual academic text production, tracing the different levels in which language professionals interact with texts.Within her section on the author's editor as wordsmith, Tacker (p. 5) establishes three types of wordface text shaping-technical editing, substantive editing, and creative editing-with one level overlapping the next.Traditionally, author's editors have worked in-house in institutional settings such as universities and other research institutes, hospitals, corporations, national and international agencies, and other such brick-and-mortar contexts, although the trend in recent years has been for these professionals to work on a freelance basis, offering their services only to those who can afford them [16].
With so much at stake in English-medium academic text production and considering the great lengths to which NNES authors have to go to publish their research, some scholars of global academic writing have begun to question the fairness and viability of subjecting NNES writing to criticism according to NES norms.Two groundbreaking texts published by Jenkins [17] and Seidlhofer [18] proposed a new way of understanding English as it is spoken in its multitude of contexts, "uncoupling the language from its native speakers" (p.151).Dewey [19] situates such researchers as Jenkins, Seidlhofer and others within what he calls the "transformationalist framework" of English as a Lingua Franca (ELF), that is, a school of thought that sees the impact of globalization on the English language as one that is reshaping the way in which English is spoken and, hence, the standards against which its learners should be taught and its practitioners should be judged.Academics in this vein advocate research into how English is used by speakers outside of "inner-circle" [20], that is, "periphery" [21] or, in the case of Spanish academics, "semi-periphery" [22] writers.
One of the measures proposed by Seidlhofer [18] is the creation of an ELF corpus, as doing so would provide a real description of English as it is used in today's globalized world.A corpus of academic texts like the ones produced by NNES academics would constitute a particularly rich specialized corpus [23] and could be used to analyze a number of qualitative and quantitative aspects.Some of these qualitative aspects include move structures and other level features (Leech, 1991, cited in [24]).Given that corpus data lack the communicative context that gives the data their meaning (Widdowson, 1998(Widdowson, , 2002, cited in [24]), complementing text-based analyses with qualitative data that take in the view of experts and informants [25] through ethnographic studies of the particular professional and socio-cultural circumstances in which language is produced can inform the results of concordance lines [24,26], and some studies have done just that [27].
This very ethnographic approach is taken up by Lillis andCurry [3] and Lillis [28]; the "text-histories" perspective adopted by the two researchers "views texts as only one feature of the social situation in which writing takes place" (Flowerdew, 2002 quoted in [25]), employing such methods as "direct observation, interviews, and other modes of analyzing the situational context" (ibid, p. 294).Other researchers have taken on an ethnographic approach to complement text-based observations on self-mentions [29] and to gain first-hand accounts of international publication practices [11].Indeed, the first decade of the 21st century saw a great development in ethnomethodological and ethnographic approaches [30], thus providing many precedents for study and helping to "bridge the gap between text and context" [28](p.373).
Of all the different research genres available to scientists, the most highly prestigious is the research article (RA) [31].Owing in part to the esteem enjoyed by the RA as the most desired vehicle of research output, since the late 1980s and early 1990s, the field of genre analysis has developed as an attempt to explore the characteristics of this and other research genres, looking to conceptualize the aspects of genres that set them apart from other types of text.The scholarship on genre analysis has provided applied linguists and teachers of English for Research Purposes with a description-based vocabulary that can be used to identify patterns in RA writing and better train both experienced NNES researchers and novice academics to write effectively.Additionally, from the perspective of the author's editor, more in-depth knowledge of the rhetorical makeup of RAs may help editors overcome their lack of specialist field knowledge referred to by Willey and Tamimoto [32], thereby allowing them to focus on discourse-level issues and provide added value beyond surface-level editing.
This paper provides an overview of a multi-year research project undertaken by the author, an institutional editor employed by a research institute in Madrid, Spain.As a practicing language broker and researcher of academic discourse in health-care contexts, my aim was to describe the academic writing produced by a sample of 10 NNES authors working in the institution who, on a volunteer basis, provided me with access to a range of materials generated throughout the process of attempting to publish one RA.The approach taken draws from two main analytical approaches to academic writing: genre analysis (text-focused analysis) and text histories (ethnographic analysis).The text-focused component examines the generic characteristics of the texts studied by analyzing their rhetorical structure in terms of moves and steps and then compares the findings of this analysis against some of the most established frameworks published to date, indicating the degree to which texts written by the authors studied are rhetorically "unique" or "standard".The purpose of such an analysis is to characterize the genre structure and communicative functions of NNES-authored texts at particular points in their history and thereby determine the evolution of genre patterns throughout the production of texts.The text-history component of the study seeks to trace the trajectory of texts from conception to publication or, conversely, to the moment in which they are abandoned.Using a text-oriented heuristic, special attention was paid to the generic characteristics of the different versions of each text and also to the influence of literacy brokers both before submission and during the peer-review process.Capitalizing on the "embeddedness" of my position within the institution, the contributions of pre-and post-submission literacy brokers as well as the reaction to such contributions are brought into focus.The findings reached by this mixed approach will continue to broaden the scope of ELF within written discourse in the biomedical sciences, shifting the focus of ELF research away from spoken language and toward a description of the writing produced by NNES researchers, offering insight into how this writing is received during peer review by English-medium international journals and providing a fuller perspective of the variety of figures involved in producing texts.An overview of one of the text histories is offered to exemplify the methodology used.

Research Design and Sample Selection
The research was conducted at the Health Research Institute of the Fundación Jiménez Díaz hospital, where I have worked as institutional editor since 2009.At the time the study was initiated, I had held the post for over four years, and in that period had come into contact with dozens of researchers seeking translation or editing services.The sample of researchers was chosen on a convenience basis, and before beginning the study I had worked with eight of the 10 volunteers and had had informal contact with another.For the sake of representativeness, all of the informants chosen were from different departments.
The 10 informants chosen worked in the following specialties and sub-specialties: immunology, genetics, lipid metabolism, pediatrics, nephrology, hematology, orthodontics, cardiology, neurology, and vascular pathology.At the time, eight held full-time salaried positions as researchers, physicians, or both, while the other two worked under other arrangements (one as a paid researcher with a temporary contract and the other as an unpaid doctoral student).

Text Histories
The study began in July 2013 when, during an informal interview held with all of the volunteer participants, each was informed of the aims, scope, and requirements of the study.The materials they were to give me were similar to those used by Lillis and Curry [3]; these included any and all drafts of one RA that they had previously published or abandoned, were drafting at the time, or planned to draft and complete over the coming two years.In addition, the informants were instructed to provide me with all email correspondence, submission letters, and post-submission comments from editors and reviewers for the article.My study of these drafts then consisted of first collating all of this evidence to determine chronological order and also to ascertain who was responsible for the modifications appearing in the drafts (e.g., the informant him or herself, coauthors, language brokers (myself or others), journal reviewers, etc.).

Move Analysis
An analysis of the rhetorical characteristics of each of the 10 RAs and a description of how and when these rhetorical changes took place was then undertaken so that these findings could lead to a fuller understanding of the characteristics of this sample of NNES writing, the individuals that helped shape these texts, and the dynamics inherent to the text-production process.For the text-focused component of the study, I set out to perform a move analysis on the texts that made up the text history; to make the time-consuming task of assigning rhetorical moves to each stretch of text more manageable, I focused my analysis on certain "milestone" versions.These textual milestones were chosen for of their significance for the overall process and also because they captured the contributions of certain brokers along the arc of the text, that is, pre-submission brokering done by me (the institutional editor) or by peers of the text's author who collaborated before the text was submitted and also post-submission brokering done by journal gatekeepers.The first of these was the draft sent to the institutional editor for revision, representing all work done by the informant and coauthors up to that point.The second was the version revised by the language editor (myself or another like me) using the track-changes function of Microsoft Word ® to mark passages where the editor made direct interventions in the text as well as margin comments to make more indirect observations such as calling the author's attention to problematic passages, requesting particularly close review, or confirming the editor's supposition, among others.Following this round of feedback received from the language broker, the third milestone version was the draft submitted to the first journal, thus providing documented evidence of the degree to which the informant and his or her collaborators acted upon the direct and indirect interventions made by the editor.If the informant successfully published the manuscript, the last milestone version was the text as published, either by the first journal or by another journal in case of rejection by the first.The move analyses performed on these milestone versions traced the course of the RA's rhetorical characteristics, and those moves considered to represent a rhetorically significant change with respect to the previous draft were noted, in which case the intervening drafts were consulted in order to pinpoint the circumstances surrounding this significant change. 1  Move analyses were broadly based on the 10-step guidelines appearing in Biber [34] (pp.[33][34].In particular, one widely-recognized framework was used to guide the descriptions of the abstract [35] and each of the main sections of the texts: Swales [31,36] for introductions, Nwogu [37] for methods, Williams [38] for results, and Dudley-Evans [39] for the discussion section.The first milestone version-i.e., the first draft sent for revision by the institutional editor-was studied, and for the subsequent milestone versions only those instances with clear changes in communicative purpose were considered.

Informant Interviews: Combining Text Histories and Genre
For each text history, this text-centered study resulted in a series of insights as to the trajectory that each text had followed, although due to the highly sophisticated specialist knowledge expressed in each, my analysis also raised a number of questions requiring clarification.In the absence of another coder whose judgments on the move assignments would have brought greater inter-rater reliability 1 I am very grateful to Theresa Lillis [33] for suggesting this modification to my original approach.and limited the subjectivity of the analyses [1,34,40], questions about the authors' intent were included in the semi-structured interviews described below.
Following this move analysis, all feedback received from journal editors and reviewers was studied in search of comments that appeared to be either directly or indirectly related to the linguistic quality of the manuscript.Special attention was paid to all feedback that addressed discourse-level issues such as the absence of particular rhetorical moves or criticism as to the inappropriate length of certain rhetorically-related components.Specific questions on the informants' views of this criticism were included.
Semi-structured informant interviews were held at the conclusion of the aforementioned move analyses.These interviews consisted of a set of questions posed to all participants and another series of questions related to each interviewee's documents that made up their text history.A comprehensive list of these standard questions asked in all interviews appears in Box 1.One interview was held with each informant-the first in February 2014 and the last in July 2015.Due to the time required to construct the text histories, perform the move analyses, and draft the questions for the informant interview, work on these studies was done consecutively rather than simultaneously.At the beginning of each interview, the volunteer participant was asked whether he or she wished to carry out the interview in English or in Spanish and then to explain the reason behind their decision.Each interview consisted of two parts.The first comprised text-focused questions aimed at clarifying aspects of the texts and determining the degree to which the text studied was representative of the informant's experience with publishing in English.In the second part of the interview, the author was asked questions related to his or her attitude toward English, English within their profession, the burden of writing in English, perceived advantages or disadvantages when writing for English-medium publications, and motivations for carrying out and publishing research.These talks lasted from 45 to 80 min and were recorded using a mobile-phone application and a similar application for a tablet as a back-up.Follow-up questions were posed to the informants by email if, during the transcription and final analysis of the text history, any aspects related to the interview required clarification.

Box 1.
Questions included in all semi-structured interviews.
1.You are the primary author of the text.What other authors contributed to the drafting of the text? 2. Can you summarize briefly the process of publishing the manuscript? 3. Did the manuscript undergo substantial changes from beginning to end? 4. Please describe your career and your position at this institution.5. Why do you perform research?In other words, how would your life or career be affected if you didn't do research?6. Do you believe that you reached your current position because you have produced scientific research?7. Are there explicit expectations that you will continue to research?8.When do you carry out your research?Is there time built into your work schedule for research and writing about your research?9. To what extent was this article a collaborative effort?10.To what extent is the effort to publish this article representative of your experience in publishing in English?Or, to the contrary, is it anomalous?11.Was it necessary for you to publish this in English, or could you have published the article in Spanish?12. Does publishing in English require a substantial effort from you? 13.Is it common for your papers to be rejected because of what reviewers or editors consider to be "deficient" English?14. Do you feel you understand what is expected of you by English-language journals concerning the type or quality of English in your articles?15.Do you believe you are at a disadvantage compared to your native-English-speaking colleagues when it comes to publishing? 16.Do you believe your experience in publishing this article would have been different if you were a native English speaker?How so? 17.What is your attitude toward the English language in general?And toward English within your profession?

Results and Discussion
Due to the ethnographic approach applied in this research project, the results are varied and tell the story of 10 individuals and the experiences of each of them.The methodological focus, however, served to shed light on a number of different themes affecting NNES academics in their attempt to see their work published in English.The most salient of these themes are (i) the influence of language brokers on the shaping of the text, (ii) the strategies applied by the author informants in drafting and revising their work, (iii) the degree to which the RAs changed as they made their way toward publication or were abandoned, and (iv) the attitudes expressed by the authors during informant interviews as they discussed the processes, challenges, and rewards that go into publishing their work for a global audience in the English language.An exhaustive treatment of the major and minor themes that came to light will form part of a doctoral thesis currently in preparation (Shaw, forthcoming).In this paper I will offer an account of one of these text histories, highlighting a number of issues that are central to the results of the research as a whole beyond the text history featured here.

Author Profile
This informant, whom I will call Miguel, is a senior nephrologist at the Health Research Institute of the Fundación Jiménez Díaz and is highly active outside of his work as a clinician.I first came into contact with Miguel before I began my research project, when he approached me to edit one of his manuscripts that he hoped to see published in an international journal.He is highly motivated by research endeavors, favoring clinical over basic research.In his 20-year career he has been affiliated with a number of universities, where he has taught in various graduate and undergraduate programs.During our 40-minute informant interview held in Spanish, he made it clear to me that, in his view, work with patients cannot be divorced from research and teaching, stating, "Doctors are in this profession to do three things: to treat patients, to teach, and to design new ideas", and later, "I cannot conceive of medicine without research.And I believe that any doctor who does is not a good doctor". 2 Miguel explained to me that the institution where we are both employed expects him to perform research like the RA included in my study, although publishing in the English language requires a substantial effort from him.Despite his positive attitude toward the language in general, he struggles to produce papers that meet the standards of the journals to which he sends his work.When asked whether he felt he knew what the expectations of English-medium international journals were, he said that he did not.(As we will see in the text-centered analysis, he received a substantial amount of criticism from the journals he submitted the text to, thus prolonging the publication process.)Miguel told me that while he is aware of his limitations when speaking English, he has successfully presented papers before "important [scientific] forums", and in his opinion is more confident when speaking about his field of science than when writing about it.
A number of Miguel's texts have been published in his native language, although when he believes a particular work to be more important than others, he tries to publish it in English so as to reach the widest readership possible.One of the journals he publishes in offers translation into English, thus relieving him of the great effort of writing in English; however, this was not the case with the RA studied here.When asked whether his work had ever been rejected due to deficient writing, he said that it had, and he believes that international English-medium journals often reject contributions from foreign scholars because of a perceived lack of linguistic quality.In his view, the peer-review process is often arbitrary, and the scientific quality of RAs that get published is sometimes uneven.Asked to explain why he believes the fate of submitted manuscripts often depends on the referees assigned to evaluate the text, he answered that a recommendation to accept or reject the study often hinges upon whether the reviewer knows the author, thereby suggesting conflation between linguistic foreignness 2 Like other excerpts from the informant interview included here, this is my translation from the Spanish.and lack of familiarity with foreign scholars' work.In spite of Miguel's struggles to publish in English, we will see below how his recent collaboration with academic language brokers inside of the institution has changed this situation considerably.

Trajectory of the Text
This text history was remarkable due to the high number of drafts made available to me by the author and the substantial effort it took to assemble these texts into a chronological order.Establishing chronological order with the highest possible degree of precision was central to the study because the analysis undertaken used these documents as physical evidence of how the RA changed over time.When I first met with Miguel to collect the materials for my study, he gave me over 60 versions of the text in digital format, with no clear indication as to the order in which they had been created or which of Miguel's collaborators had instigated the changes appearing in each new version.It appears that while author collected most all drafts produced by his collaborators and himself, the revisions made in earlier versions were either accepted or rejected by the author, who worked only on the most recent text and not with any previous drafts.This practice indicates that after sending the manuscript for commentary, the text always returned to its author for evaluation of these comments, thereby suggesting that the author prefers to revise these pre-submission edits rather than discuss his colleagues' proposals with other collaborators.
To overcome the obstacle posed by the number of texts and the difficulty of establishing chronological order, I ordered the texts based on the name given to each file, as many of these contained either the names or initials of his collaborators, the dates when these contributions were made, or the names of journals or abbreviations thereof.When this criterion was insufficient due to repeated file names or conflict between the information given in these names, I consulted the "last modified" date in the document properties.This method I used to empirically piece together the texts in the correct order introduced a certain margin of error.This methodological shortcoming may have come about because of my failure to convey the importance of chronological accuracy, and future studies adopting this method should take greater care in this regard.Despite this unexpected difficulty, during the analysis I found that the changes made between drafts-both milestone and non-milestone-did not suggest that this shortcoming had had a substantial effect on the validity of my findings.
Miguel reported drafting this manuscript in collaboration with several colleagues, with the collaborators choosing to write the text partially in English but mostly in Spanish for translation at a later time.The chronology of the text shows that the first 19 drafts were in this mixture of the two languages, and then the remaining versions are in English.For the purposes of my study, I included only the last of these texts written partially in Spanish in order to analyze the effect of the transition between the two languages on the rhetorical characteristics of the texts.This text history contained four milestone drafts: the version sent for pre-submission revision by the language broker, the changes proposed by the language broker, the draft as submitted to the first journal, and the final published text.
As detailed below, the text was met with considerable resistance by the all three of the journals it was submitted to for consideration, although the third eventually agreed to publish it after substantial revision.Issues with the quality of the language were raised, although Miguel believes that the central message of the text remained intact from start to finish.

Collaborators and Pre-Submission Language Brokering
During the early stages of drafting the text, the author asked four different colleagues of his-all native Spanish speakers who are members of his disciplinary community-to comment on the manuscript, and the contributions of these individuals were more focused on what was being said in the paper rather than how these messages were getting across.Evidence of these exchanges appears in a number of drafts included in the text history as margin comments in Spanish.This feedback took two different forms: on the one hand, text written in Spanish to be translated and then inserted into particular passages, and on the other, more indirect commentary indicating how the collaborator suggested altering the text though without phrasing these suggestions.An example of this first type of comment is seen in the discussion section of the third version, which was the last draft created before the document was sent to me for editing.The collaborator writes a comment in Spanish which in the following draft was translated as "Few [hemodialysis] units measure the serum calcium at the beginning and after the dialysis session".Later, in the sixth draft (between milestone version two and three), a colleague highlights an entire paragraph of the discussion and writes in the margin: Careful with this! Postdialysis increase of proteins sequesters Ca++ and to compensate [sic] and so that Ca++ remains stable (since it is controlled by PTH) total calcium is increased (this is why reduced calcium cannot be evidenced with a 2.5 bath when evaluating total calcium).When 48 hours pass and proteins have normalized, Ca++ remains low (although not as low as in post-dialysis) and total calcium also decreases (through not significantly).
In spite of the agrammatical nature of this scientific shorthand, what is clear is that the coauthor here calls Miguel's attention to possible inaccuracies in the assumptions he introduced in the text.
Of the 10 text histories included in my research project, this was one of nine in which I was a collaborator.My first hybrid translation/edit is fourth in the text history, and contains both direct interventions in the text marked using the track-changes function and margin comments in which I called the author's attention to other aspects of the manuscript.As with many other academic texts sent to me for editing, the RA contained a number of passages that appeared to require substantial reworking for the sake of clarity.Most of my suggested changes to the manuscript were recommendations for clarification due to syntactical errors, while others included requests to improve transitions between ideas, style-based remarks in which I advised against one-sentence paragraphs, and also some advice on adapting the text to the target journal's instructions for authors.An example of both direct and indirect interventions can be seen in the following two versions of the same passage in which I suggested that the author modify certain wording to make it more formal and also pointed out that his argument could be improved by further elaborating on the significance of his references to current knowledge in the field: Original text: At present, there is much debate about using dialysis with low calcium concentrations.Both the KDOQI xiii as the KDIGO 11 guidelines recommend using low calcium concentrations in the dialysate.
Revised text: At present, there is much debate surrounding dialysis with low calcium concentrations.Both the KDOQI 14 as well as the KDIGO 11 guidelines recommend using low calcium concentrations in the dialysate.
In suggesting a change from "about" to "surrounding" and correcting "as" in favor of "as well as", I displayed whatTacker [15] refers to as "technical editing", but also included the following comment in the margin: "This should be followed by a comment on the relevance of these guidelines or whether the data or your argument support such a position."This second intervention of mine can be accurately described according to Tacker's typology as "substantive editing" because in addition to a focus on surface-level features, I recommended including explicit mention of where the author situates his view.Though one of many such interactions with the text, this excerpt from the first milestone version supports Tacker's assertion that in practice texts require author's editors to blend the three types of intervention as they work to improve NNES-authored texts.
Days after I returned the manuscript to Miguel with my proposals for improvement, he surprised me when he asked if I would like to be included as one of the article's co-authors, and I was delighted to accept his gracious offer.As a result, my name appears in the drafts included in versions 8 through 17.As of draft 17, however, my name was removed from the list when Miguel was advised by one of his eight coauthors that editors are not customarily included as coauthors.When I asked about this gesture and his subsequent change of mind, Miguel explained that he believes all contributors merit recognition, and that he felt that if it was considered proper to include his statistician as a coauthor, then a language professional like myself should be equally regarded.Granting authorship to language professionals is misguided, however, as they do not meet the criteria to be regarded as such [41].First, editors normally play no part in the conception or design of research studies or the collection and interpretation of data, thus disqualifying them as potential coauthors.Second, while manuscripts may undergo substantial change during the editing process, most editors are neither prepared to assess the scientific merit of a manuscript nor tasked with doing so [14].Lastly, authors have no obligation to share the final version of their texts to the editor, and in my personal experience all proposed changes to manuscripts are mere suggestions to be taken into consideration by the coauthors.Based on these criteria, the author was wise to accept the counsel of his peer and eliminate my name from the by-line.
At the beginning of each informant interview, I asked the volunteer participants to indicate which of the named coauthors appearing in the final published RA had played a role in writing the text.In particular, I asked the informants to make the distinction between individuals whose input was aimed more at issues of content and those who had offered ideas that had implications for not just what was being written about but how that information was portrayed.Miguel answered that one of his collaborators had made a substantial contribution to both content and rhetorical structure, giving what Miguel called a "critical appraisal" of the language used in the text and also the way the text had been developed.This individual, whose initials appear on drafts 19 to 21, not only advised Miguel in reworking the RA but also gave him advice on how to orient his point-by-point letter in response to reviewer feedback.A seasoned veteran of Miguel's disciplinary community, editor-in-chief of a prestigious international scientific journal, and a fellow employee of the institution, this collaborator's influence was decisive to the fate of the article, acting at once as a pre-submission peer reviewer and literacy coach, that is, a high-value academic literacy broker.In section 3.3 below, I will go into some detail on the concrete modifications that helped make the text more acceptable for the final journal, though in this section I will recount some of the information I gathered during the informant interview on what has made this individual so adept at publishing internationally.
I was unable to access documentary evidence on the specific recommendations made by the academic literacy broker, and it is likely that much of this insight was transmitted verbally between Miguel and this colleague.From Miguel's account of the process, however, it seems that this individual contributed to the text in a way that can accurately be described as "creative editing" according to the framework of Tacker [15].After having described in the early part of the interview the role played by this collaborator, Miguel invoked this individual once again when I asked him if he believed his own personal and professional ties with researchers in Anglophone countries may have helped Miguel publish this paper.He stated that such connections were not a factor in this RA, and that if he did have influential figures on his side, the endeavor of publishing texts like this one would be much easier.To this he added that he could probably count the number of scientists who had earned this kind of influence among journal editors and reviewers "on one hand", and that the academic literacy broker who had assisted him with his publication may well be one of them.In Miguel's view, the academic literacy broker has reached the position he currently holds because of his proficiency in writing about his findings thanks to a fellowship in the United States and because he is so well-read in the field of science.Additionally, Miguel underscored his colleague's resourcefulness when navigating the genres and dynamics inherent to their field, thus suggesting that this individual's great literacy is not only a matter of knowing what to say or write in particular circumstances, but also how to act in given professional situations.As a result of this individual's contributions to this text and others that have followed it, today Miguel believes that his "problem" with publishing in English "has gone away".

Journal Response and Post-Submission Language Brokering
The only reviewer feedback I was able to access was from the third and final journal.Two reviewers provided feedback on Miguel's paper.Of the thirteen comments made between these two reviewers, the first reviewer remarked on two aspects that changed the rhetorical shape of the final product, while all of the second reviewer's comments concerned issues related to content rather than the way the text had been written.Both of these sets of comments contain a number of errors in syntax and morphology, indicating that they were either mere errata or that the reviewers themselves are NNESs.
The first comment made by reviewer 1 is as follows: "Title: Considering the results I would recommend to include [sic] the word 'individualization' in the tittle [sic]."Miguel had referred to individualization on three occasions, stressing that relevance of his findings pointed to a need to individualize calcium dialysate levels based on a number of factors inherent to each patient.Heeding the reviewer's advice, Miguel changed the title of his text accordingly, though as we will see in the text-focused analysis below, this input from the academic literacy broker led Miguel to revise the body of the text as well, amplifying the effect of this feedback on the final product.
The second of these comments urged Miguel to revise the discussion, claiming it was "too long and repetitive".Asked to respond to this criticism, he stated with some resignation that reviewer comments like this one are tantamount to demands because of the status of journal gatekeepers.Miguel conceded that the manuscript as a whole may have been abnormally long for an English-medium publication, although he also defended the scientific merit of his method of writing the discussion, stating that the text he felt obliged to remove served to "support the paper's arguments by citing the existing literature and then contrasting these findings" from different angles.The final, published version of the article shows that Miguel's revised text is more synthetic and less devoted to listing findings, and the length reduction he made was deemed sufficient for the article to be accepted.The author attributed this supposed shortcoming in his style to differences between academic texts in Spanish and English, stating that English texts require greater conciseness.In this regard, we see how Miguel's rhetorically "foreign" style hindered his effort to publish the RA, thus suggesting that this instance of ELF in written discourse was taken to be inappropriate for the target journal.

Genre Considerations
Due in large part to the input received from his coauthors, my proposed changes to the manuscript, the high-value academic brokering received from his colleague, and the feedback provided by journal gatekeepers, this RA went through a somewhat substantial reworking of its rhetorical characteristics between the time it was conceived and its final published version.As I learned by studying the history of the text, a number of factors seem to have led Miguel's text down a somewhat circuitous path.These include such issues as code-switching practices Miguel uses when writing scientific texts that will eventually be published in English, efforts to sharpen the text's focus while situating the study within the context of the ongoing scholarship within the field, and recasting of results in light of reviewer comments.Below I will examine some of the most significant rhetorical modifications made to the text and complement these insights with information gleaned from my study of the text's trajectory.
One of the more noteworthy dynamics that came to light as I studied the development of the text throughout its many drafts and sources of input appeared in the transition between the drafts mostly written in Spanish and those written in English.Of the more than 60 versions the author provided me with, the first 19 of these are mostly in Spanish, although as of the twentieth he worked almost entirely in English.When I performed a move analysis of each of the paper's main sections and the abstract, it became clear that the versions of the RA in the two languages were entirely the same, thus raising the question of whether he understood genre structure to be the same in the two languages or if other factors were at play behind this strategy and also hinting that, at least at this particular stage, the text was a pure form of written ELF in that the initial rhetorical structuring of the paper in English mirrored the text's rhetorical makeup when it had been written in Spanish.I inquired about this rhetorical sameness between the Spanish and English versions, asking Miguel why he worked this way.He was unable to think of a reason for this other than the fact that he has always used this approach, although to my follow-up question of whether academic texts are written in the same way between the two languages, he quickly replied that they are not, adding that texts in English that are "Spanish-ized" are clearly inappropriate.In his view, there are particular "turns of phrase" and "expressions" that are particular to English texts, and learning to master this way of writing is exceedingly difficult.On this same topic, Miguel also mentioned that Spanish sentences tend to be longer, while in English there is a premium on conciseness.I gathered from Miguel's response that he believes the differences to be articulated at the sentence or paragraph level rather than at the level of discourse.However, given the fact that many of his collaborators who contributed to earlier drafts were focused on content-level aspects of the paper and provided input in Spanish, there may be more dynamics at play behind this code-switching.
One significant theme that came to light across the different drafts of the text is a narrowing of the text's focus, paring down lengthy stretches of text devoted primarily to one rhetorical move.This effort was largely centered on introductory or background moves such as Swales's Move 1 (establishing a territory) in the introduction [31] and what Dudley-Evans calls the "information move" (Move 1, where authors present background information about theory, research aims, methodology used, and relevant previous research) [39].Evidence of this adaptation is particularly present in the introduction section, which was revised substantially as the text made its way toward final publication, and especially in the post-submission stage.We can see how this happens by comparing a fragment of the introduction in the draft sent to the first journal (i.e., milestone draft 2) and the final published text (milestone 4).As seen in the extract appearing in Box 2, the revised text provides only the general thrust of this background information, while the original draft goes into some degree of detail on the recommended calcium intake for hemodialysis patients and how other studies in the literature have established these optimum levels for patients with different pathologies.(a) The recommended daily dose of calcium according to the Institute of Medicine's recent report on dietary reference intakes is 1000-1200 mg/day, which is the amount needed to meet the requirements of 97.5% of healthy adults (4).Stage III-IV chronic kidney disease (CKD) patients show decreased urinary excretion of calcium (100-150 mg/day) while maintaining normal fecal excretion.In these patients calcium balance is neutral with an intake of 800 mg of calcium, and very positive at 2000 mg/day.However, these results can change depending on whether there is supplementation of vitamin D or not.In CKD III-IV patients not treated with vitamin D, the intake recommended to obtain neutral balance would still be clearly positive with an intake of 800 mg/day of elemental calcium, and 800 mg/day in patients who do receive vitamin D treatment (5).In anuric patients with no urinary calcium loss, the balance would still be clearly positive with an intake of 800 mg/day.Calcium balance is complicated when the patient is on hemodialysis (HD).If a negative balance of 200-300 mg/session is possible, the patient will continue with a positive balance with an oral intake of 800 mg of calcium/day (4).
(b) The recommended daily oral calcium intake according to the Institute of Medicine is 1000-12,000 mg/day, which is the amount needed to meet the requirements of 97.5% of healthy adults (4).Urinary calcium excretion contributes to calcium balance.Lack of this adaptive mechanism in anuric chronic kidney disease (CKD) patients may result in a positive calcium balance from the intake of 800 mg calcium/day (5).Guidelines currently recommend reducing oral calcium intake, including calcium-based phosphate binders in hemodialysis patients, especially in the presence of adynamic bone disease, vascular calcification or hypercalcemia (6), to decrease the risk of vascular calcification (7).In this regard, over 60% of dialysis patients have adynamic bone disease (8).Dialysis calcium concentration ([calcium]) also contributes to calcium balance in hemodialysis patients (9,10).The choice of dialysate calcium concentration ([calcium]) will determine calcium balance during hemodialysis and, eventually, vascular calcification (11,12).There is an ongoing debate surrounding the ideal dialysate [calcium] concentrations.

Box 2. Cont.
Two types of exogenous calcium are administered to HD patients: on the one hand, oral delivery, which depends on diet and calcium phosphate binders, and is regulated partially by vitamin D treatment; (6) and on the other, input through calcium dialysate (7).Guidelines currently recommend reducing oral calcium intake, including calcium phosphate binders in HD patients, fundamentally in patients with adynamic bone disease and/or vascular calcification and/ofhypercalcemia, (8) as evidence demonstrates that oral calcium administration leads to increased calcium in the aorta.(9) Regardless of their serum PTH levels, over 60% of dialysis patients have adynamic bone disease, so knowledge of the patient's calcium balance in HD is of great importance (10).An appropriate concentration of dialysate calcium is crucial in the management of HD patients.The concentration of calcium in the dialysate is as important as the oral calcium intake, and must be included to explain the causes of calcification in HD (11,12) At present, there is much debate surrounding dialysis with low calcium concentrations.( . . . ) In this fashion, the apparently more "acceptable" rhetorical fashioning of the introduction moves more quickly past Move 1 in order to arrive at Move 2 (establishing a niche).This criticism from the journal reviewer once again indicates a certain unwillingness to accept rhetorical styles that represent a written ELF.In our interview, Miguel expressed regret over what had been left out of this more concise version, though he also conceded that readers requiring greater depth can refer to the sources listed in the paper's references.Aside from this transition away from lengthy background moves, while it is true that peer reviewer 1 urged Miguel to shorten the text and avoid repetition, these remarks were directed not at the introduction but at the discussion, thereby suggesting a "ripple effect" that peer comments can have across an entire RA independently of the specific focus of the feedback.
While the methods section received comparatively little attention by the authors and peer reviewers, remaining virtually unchanged throughout the text history, the results section underwent substantial change.The most salient differences appear between the first and second milestone drafts, with Miguel opting to include discourse markers such as "nevertheless" and "on the other hand", and intensifying adjectives such as "important" to describe differences in the values he observed in his study.In this fashion, the author brought his rhetorical intent into sharper focus though without making actual changes to moves or move cycles.
In what appears to reflect clear influence of the journal gatekeepers' feedback on the results section, the final draft introduces a number of instances of what Williams [38] calls Move 0.3 (indicating procedural issues related to the experiment), marking the transition between the author's account of the first component of the study (i.e., measuring the effect of two different dialysate concentrations on a number of parameters relative to patients' baseline measurements) and the cutoff points nephrologists may use when individualizing hemodialysis treatment.In this revised approach, Miguel adds explicit mention of the study's end goal of treatment individualization, apparently in response to the reviewer comment that read, "Title: Considering the results I would reccomend [sic] to include [sic] the word 'individualization' in the tittle [sic]".The author not only changed the previous title from "Importance of basal serum calcium in the optimization of dialysate calcium concentration" to "Individualization of dialysate calcium concentration according to baseline predialysis serum calcium", but also added a number of references to this notion, with some realizations referring to the steps they took to carry out the experiment, and others addressing the reasons behind these measures.In addition to these moves, the final published text adds five instances of Move 3.4, which indicates the implications of the study's findings.
As mentioned above, one of the reviewers made specific reference to the length and repetitiveness of the discussion section.This post-submission feedback led the author to make a number of changes to the section, as evidenced in Box 3, which includes excerpts from milestone version 3 (version sent to the first journal) and milestone version 4 (final published text): (a) Especially interesting is the fact that serum calcium and PTH levels returned to baseline values 48 h later and prior to the next HD.In humans it is possible to measure serum ionic calcium content, but there is no possibility of measuring mass balance or total calcium.Although it is an instructive concept, it is not enough to measure serum ionic calcium because it does not consider the movement of calcium from one compartment to another in the body.The fact that calcium and PTH values return to baseline before the next HD session suggests a redistribution of serum calcium and change of compartments.
If there has been a positive balance of calcium in anuric patients, obviously the normalization of the increased postdialysis calcium serum to baseline values implies a movement from the blood compartment to another undisclosed located and that could be the extracellular space and could form part of the soft tissues and/or vascular calcification.
(b) The fact that calcium and PTH values returned to baseline before the next hemodialysis session suggests the redistribution of serum calcium to other compartments, such as bone, soft tissues or the intracellular compartment, especially in anuric patients.Vascular calcification is prevalent in CKD patients and is associated to cardiovascular events and mortality (18,19,20,21).Imbalances between factors that promote and factors that inhibit calcification interact with a positive balance or high concentrations of calcium phosphate to promote vascular calcification (22,23,24).
When reducing the length and repetitiveness of this fragment, the author eliminates reference to certain concepts which appear to be understood by the reader, and the final version gets these ideas across in a way that is more succinct.

Conclusions
A number of conclusions can be reached based on the text history presented here.First, the author's use of written code-switching appears to have influenced the text's rhetorical makeup, and the accounts gathered in the text history suggest that this use of rhetorical ELF may have contributed to the negative feedback received from the journal gatekeepers.Future studies into code-switching in the writing processes of NNES academics could illuminate best practices, expanding knowledge of effective strategies for non-Anglophone academics throughout a text's often long arc and continuing exploration into the rhetorical acceptability of ELF by Anglophone journals in biomedical RAs and other genres.Second, the rhetorical changes made to the this RA-including narrowing the focus of moves presenting background information and making explicit reference to certain methodological steps in an effort to more effectively signal shifts between different components of an experiment-demonstrate certain initial shortcomings in the rhetorical structure in the text, offering specific evidence of ELF writing and the faults found with this writing by journal reviewers and editors.As evidenced here, it appears that NNES authors may sometimes be urged to refrain from including related yet potentially superfluous commentary that can be understood as causing the text to repeat the same concepts while adding little to the conceptual thrust of the paper.Another finding of interest involves the ways in which reviewer feedback may have a ripple effect across texts' generic characteristics, leading the author to alter the focus of their paper or, at least, foreground ideas already present in the previously submitted draft.Last, the pivotal role played by the in-house academic literacy broker in shaping the text suggests that these high-value brokers combine linguistic aptitude, expert knowledge of the disciplinary community, name recognition, and scientific expertise, and the value of this multi-faceted academic literacy can complement the more modest contributions made by language brokers focusing more on surface-level issues with texts.

Box 2 .
(a) Fragment of introduction section as sent to the first journal (text conserved in final version marked in bold); (b) Final, reduced version of the same fragment.

Box 3 .
(a) Fragment of discussion section as sent to the first journal (text conserved in final version marked in bold); (b) Final, reduced version of the same fragment.