‘I Have to Hold Myself Back from Getting into All That’: Investigating Ethical Issues Associated with the Proofreading of Student Writing

L2 student writers in UK universities often seek the services of a ‘proofreader’ before submitting work for assessment, and the proliferation of freelance proofreaders and online proofreading agencies has led to debates about the ethics of the proofreading of student writing in publications such as Times Higher Education. This study investigates the kinds of ethical issues that confront proofreaders and how they deal with them. Nine UK university proofreaders proofread the same poor-quality L2 applied linguistics master’s essay, explaining their interventions by talking aloud while proofreading and at a subsequent interview. Proofreaders addressed ethical difficulties by means of two macrostrategies: (i) selective proofreading; and (ii) declining to proofread in part or in whole. Two additional findings relating to ethics emerged from the study: firstly, some informants experienced dilemmas and uncertainties despite their attempts to proofread ethically; secondly, a number of informants went far beyond traditional, narrow conceptualizations of proofreading, making interventions affecting the writer’s structure, argumentation, and content which could be seen as unethical. The findings highlight the need for the regulation of proofreading to ensure it is standardized and consistently administered from writer to writer, and I close by recommending that universities strive to implement more formative types of proofreading to enhance writers’ academic literacy, not just their texts.


Introduction
It is easy for UK university students to find a 'proofreader' for their essay via adverts and flyers stuck on campus walls, or by contacting an online agency. Indeed, Harwood et al. (2009)

Arguments for and Against Proofreading as Ethical Intervention
McNally and Kooyman (2017) summarize various pro-and anti-proofreading arguments made with reference to ethics. Anti-proofreading arguments focus 'on integrity, authorial ownership and the development of student writing competence, and plagiarism' (p. A-147). There is a danger that the proofreader may rewrite a text to a standard beyond the capability of the writer. As a result, …the final reader/marker, unaware of the nature and degree of external intervention in the final version, will award a grade that does not accurately reflect the student writer's real ability. (p. A-147) The proofreader could also enhance a writer's argumentation, raising the possibility of appropriation. Composition scholars (Brannon and Knoblauch 1982;Sommers 1982;Sperling and Freedman 1987) discuss how students' writing teachers may appropriate the text by means of their feedback, but the concept can easily be applied to proofreading: student writers may revise their texts to make meaning-level changes the proofreader rather than the student thinks are necessary. Again then, the marker would be grading the proofreader's rather than the student's text. Other unethical forms of proofreading are much discussed in outlets like The Guardian's higher education blog and Times Higher Education. For instance, Molinari (2014) highlights the unethical nature of some online proofreading agencies: If you Google Bacademic proofreading,^you will see a list of sites offering to Bproofŷ our work. What they are also offering, however, is to write your assignments for you.
Molinari bases this claim on two kinds of experience. The first is of occasionally spotting ghostwritten work when assessing students' essays, when a writer 'submit[s] a piece of beautifully polished and referenced work that is clearly at odds with evidence from our dayto-day interactions [with the student].' The second experience concerns her knowledge of unscrupulous proofreading agencies. She describes how 'a fit of professional curiosity' led her to pretend to seek work with one such agency and how, during her interview, …I was asked if I would actually write essays [for clients] and told that I would be paid according to whether the essay got a first or a second.
Returning to McNally and Kooyman's discussion, another anti-proofreading argument suggests that the writer may not reflect upon the changes made before submitting the work for assessment. If the proofreader has used Word's Track Changes tool, the writer can simply accept all amendments without noticing and remediating frequent errors. Hence proofreading is unethical since it militates against writer development, the goal of alternative forms of English language support such as writing centres, which strive to enhance the writer's knowledge rather than quickly fix the text (see North 1984). A final anti-proofreading argument concerns disagreements around ethically legitimate forms of proofreading; there is the potential for inconsistent and unjust treatment of writers without precise definitions in place. As McNally and Kooyman put it, Without agreed guidelines…, the provision of proofreading services is fraught with difficulties regarding inconsistency of feedback and a blurring of the line between Bsurface^and Bsubstantial^corrections, which could result in a plethora of damaging three-way disputes between students, faculty, and [language] advisors. (p.A-147) McNally and Kooyman have in mind in-house proofreading offered by language tutors; but of course the same problems could arise if freelance proofreaders were offering different degrees of intervention.
In marshalling pro-proofreading arguments, McNally and Kooyman point out that while language professionals' long-term goal is to develop writers, the 'short-term concern of the student…is for concrete support for their paper right now ' (p.A-148). Furthermore, it is unrealistic to presuppose that every student can cope with written assignments, unlike the 'homogeneous university system of yore where a basic level of written literacy could be assumed ' (p.A-148). Whatever we may think of the wisdom of current (low) admissions policies (often driven by a reduction of government funding, leading to universities being willing to take a lower standard of applicant than previously: see Hadley 2015), it is 'morally incumbent' (p.A-148) on universities to provide support for these struggling students-and proofreading can play its part. Moreover, ethical proofreading can be pedagogically focused: 'Proofreading as part of a collaborative process can reveal gaps and weaknesses in the writer's grammar and syntax, and offer Bteachable moments^and learning opportunities that provide valuable individualised feedback' (p.A-149). In the same vein, Bowers (2013) attests to proofreading's pedagogic potential. Bowers describes a type of proofreading different from 'laborious line-by-line corrections of every aspect of grammar, spelling, and syntax' which does writers 'a disservice'. Instead, the proofreading he advocates is used 'to show examples of where [writers] are making errors based on a sample of their writing, and to explain how to identify and correct similar mistakes in the rest of their work. By modelling good practice, we aim to improve their selfefficacy in the future'. In sum, then, although many accounts in the educational press focus on less scrupulous forms of proofreading, it has its defenders, who suggest that implemented properly, it can be a highly ethical, developmental form of intervention.
The foregoing discussion gives a flavour of the heated debates on the ethics of proofreading. But much thinner on the ground than discussions about proofreading is empirical research focusing on the proofreading of student writing in general or the ethics of proofreading in particular. 4 I now focus on two such studies.
Research on the Ethics of Proofreading Lines (2016) provided two sources of evidence to support her claim that there are many proofreaders offering illegitimate and unethical forms of help. Firstly, she cited her experiences of running a large proofreading business in Australia, when she was approached 'on many occasions' (p.376) by students wishing her to ghostwrite compilations of work they had cut and pasted into a document from the Internet. Secondly, Lines posed as an L2 PhD student and contacted 50 proofreading services to ask whether they would engage in unethical proofreading, helping her to improve the ideas and argument and correct inaccurate factual information in her draft thesis. 44 of the 50 services contacted agreed to oblige. Harwood et al. (2009Harwood et al. ( , 2010Harwood et al. ( , 2012 interviewed 16 proofreaders of student writing at a UK university, their first article offering an overview of their findings related to the proofreaders' profiles, beliefs, practices, and experiences. The extent to which proofreaders reportedly intervened varied, with some proofreaders prepared to go further than others. While most informants claimed to confine themselves to correcting 'grammar, syntax, spelling, stylistic errors' (p.179), one proofreader, Eve, would rewrite 'inelegant (but grammatically accurate) prose' (p.176) and comment on writers' ideas, a level of intervention deemed by most other proofreaders to be ethically off-limits. However, some proofreaders reported that it was not always easy to determine the cut-off point between ethical and unethical interventions. Another informant, for instance, wondered whether she would alert the writer if they had failed to address the essay question and were going off-topic. Harwood et al.'s (2010) second article focused squarely on ethical issues. Proofreaders spoke of the types of texts that it was ethically (un)acceptable to correct, and low-proficiency texts were seen as problematic because of the work needed to get the text up to a submittable standard. As one of the proofreaders, Tom, explained, 'correcting 'two cat was sitting on the mat' is not a terrible thing'; but 'when you have to work out how many cats were there in the first place, were they sitting or were they in fact standing', ethical issues came to the fore, and Tom's policy was to decline to work on such texts (p.58). Other informants debated whether they would highlight questionable arguments or 'factual errors' (p.59) because they worried that doing so went beyond their remit. One proofreader, Chloe, was willing to make these more substantial argumentation/content interventions for other types of texts (e.g., conference abstracts, manuscripts to be submitted for publication), but refused to do so for student essays on ethical grounds. However, many informants were prepared to go further when commenting than correcting. It would then be the writer's responsibility to address the proofreader's suggestion by rewriting, compared to when proofreaders simply made a correction themselves, leaving open the possibility of a passive, unreflective acceptance of the correction on the part of the writer. Hence one informant, Stella, alerted writers to what she saw as an over-reliance on quotes rather than rephrasing them herself. Disturbingly and in agreement with Lines' (2016) findings, 14 of the 16 informants had received ethically inappropriate requests from writers: for example, requests to ghostwrite or to write an essay based upon the student's notes. Like McNally and Kooyman, many of Harwood et al.'s informants felt ethical dilemmas could be resolved, at least in part, if universities produced clear proofreading guidelines; without them, some complained of feeling unsupported and 'working in the dark' (p.63). Such guidelines could then educate writers as to appropriate expectations of the proofreader.
Harwood et al.'s studies were wholly interview-based, and the authors concede that a stronger research design which featured actual rather than reported behaviour would provide more robust evidence of what proofreaders do. They suggest having informants proofread the same text, to enable a more direct and valid comparison. They also suggest having informants talk aloud while proofreading and interviewing them afterwards, in order to uncover the reasons they proofread in the way they do. The current research incorporated these methodological suggestions. As in Harwood et al.'s work, ethics again emerged as a major theme in the present dataset, and, having elsewhere provided a detailed quantitative analysis of the frequency of corrections and comments they made (Harwood 2018), I focus in this article on the theme of ethics, and on the following research questions in particular: -What ethical issues arose for proofreaders? -What strategies did proofreaders use to deal with ethical issues? Why? I now describe the design of my study.

Method
Four methods were used to collect data: (i) a proofreader profile questionnaire; (ii) text analysis of the interventions informants made while proofreading an authentic master's-level essay written by a Chinese learner studying applied linguistics, which had been awarded a bare pass mark of 50; (iii) a talk aloud task, proofreaders explaining aloud what they were doing and why as they proofread; and (iv) a post-proofreading interview. In the method section of Harwood (2018), I provide a detailed description of each instrument, and in what follows, I focus particularly on how ethical issues were explored and uncovered.

Proofreading Task
All 14 proofreaders worked on the same essay so that their interventions could be compared, as well as their difficulties and dilemmas, often ethical in nature. The essay is reproduced in Appendix A, and it can be seen that the text suffers from major language and content flaws. To give a flavour of the essay, here is an unredacted excerpt where the writer is reviewing the findings of an earlier study: In order to support Truscott's view, he design an experiment (2008): forty-seven students were divided into two groups-experimental group and control group-to be enrolled in three sections of a writing seminar. After the first article, experimental group received their article with errors underlined and need to revise their article, but controlled group received no-marked draft. It is easily to see that experimental group performed better than control group on revisions and error feedback made a positive effect on students' rewrite. One week latter, students had a new article. Compared error rate of the second article with the first one, both group received the equal results, that is to say, there is no positive effort on experimental group second article although they have their first article revised and rewritten. There are some limitations in this experiment: firstly, students could gain knowledge in a short time; secondly, the Truscott's experiment only focus on one type of feedback. At last, Truscott draw a conclusion that there are no evidence to show the effectiveness and efficiency on improving students' writing ability by correcting errors.
Various details about the study being reviewed are unclear: what is meant by 'three sections of a writing seminar'? Is the meaning of 'no-marked draft' sufficiently clear? The writer claims it 'is easily to see' the difference in the results between the two groups after the experiment-but provides no concrete details of what these differences were. The limitations associated with the study design also lack clarity: what is meant by 'students could gain knowledge in a short time'? Or that the research focused on 'only one type of feedback'? There are additionally many language problems, not least with mastery of tenses (he design; need), with citing conventions (an experiment (2008)), with the English article system (better than control group; the Truscott's experiment), with fixed phrases and collocations (it is easily to see; made a[n] effect), with number (both group), with confused words (effort vs. effect in no positive effort), with genitives (experimental group second article), with prepositions (limitations in, on improving), with agreement (experiment only focus on), with connectors (At last), and with uncountable nouns (are no evidence). Hence informants could have made various language and content interventions, depending on their conceptualization of ethical proofreading. Informants chose to proofread on paper or electronically as a laptop computer with Internet connection was provided, additionally enabling them to consult online resources (e.g., dictionaries) to aid their proofreading. The lecturer's essay brief and the department's handbook which contained information about the preferred author-date referencing style was also available. Informants were instructed to 'do whatever you do normally when you proofread'. The university at which this research was undertaken had a proofreading policy in place and publicly available, but, as discussed earlier, I was conscious of the possibility that informants may not be aware of it-or even if they were, routinely ignored it. Hence my request to proofread as they would normally. The policy issue is revisited when the results are discussed below.

Talk Aloud and Post-Proofreading Interview
As they proofread the text, informants verbalized what they were thinking and why they were making/declining to make their interventions. 5 As they did so, I made notes on those remarks I wanted to learn more about at interview, such as the following comments touching on ethical issues uttered by various proofreaders: -I would write it differently, but I can't change much. -The writer can make that decision herself. -…if the student can't even make it clear what she wants to say there's not much I can do. But I'll have a go. -I'm not willing to put more effort into this because the student doesn't seem to understand very well. This is not my business.
-This text isn't worth proofreading! -I'm going to keep it as close to the tense she is using. -I could spend ages trying to work out what she means, but I think it's up to her. So I'll just write 'What?' -[Contemplating whether to correct definite/indefinite article errors] I'm not systematic. I have to make a call about whether the article error will annoy the lecturer or not.
-In my email [covering email to the writer], I'd say I can't do a perfect job on this because the English is so poor.
Rather than waiting until a writing task has been completed and questioning the writer about his/her actions and processes retrospectively (with the attendant problems of limited informant recall of their thoughts and motivations for their actions: see Greene and Higgins 1994;Hammersley and Gomm 2008;Prior 2004;Smagorinsky 1989), talk alouds provide data about the writing process (or in our case, the proofreading process) concurrently. Although this concurrent data will not be comprehensive, as a talk aloud cannot capture the informant's decision making processes in full, it will likely provide more comprehensive data than relying solely on other methods, such as a post-task interview (Hayes and Flower 1983). For my purposes, then, the talk aloud method was able to offer insights into proofreaders' thinking and decision making where ethical issues arose. I was able to note down the relevant excerpt from the informant's talk aloud as well as the part of the student's essay which was associated with these difficulties, then question the informant further at the post-proofreading interview, rather than being obliged to depend upon decontextualized questions about their practices in general. Although I audiorecorded the talk alouds, I transcribed excerpts from informants' talk alouds relating to ethical issues as they proofread by listening and observing as the proofreaders worked on the task. This was because the post-task interviews followed immediately after the completion of the proofreading/talk aloud to lessen informant recall issues. The audiorecorded talk alouds were used subsequently to check the accuracy of my transcriptions.
I now give more details about the interview. At interview, as well as asking follow-up questions about informants' profiles which they had written about in the preliminary questionnaire, I also asked some generic questions potentially connected to ethical issues (Did you find anything particularly problematic about the proofreading task?), questions about how informants saw the proofreader's role (which at times stimulated discussion of ethical/ unethical roles proofreaders may be asked to play by writers), and tailored questions about informants' talk aloud utterances. These tailored questions included asking about specific interventions informants made/declined to make, such as those focused on the writer's argumentation and content (e.g., You commented that the writer should provide more details of the limitations of the study. Could you say more about that?). In addition, there were questions about some of the proofreaders' actions I had observed while they were proofreading, e.g., searching online for the source texts the writer was reviewing in her essay, this question helping me understand how far proofreaders were prepared to go to understand the writer's argument. An example interview schedule is included in Appendix B.
An in-house research grant enabled me to pay informants a fee of £40-plus travel expenses. Informants were recruited by contacting proofreaders known to me or to my colleagues; and by contacting proofreaders unknown to me who advertised their services around the UK research-intensive university research site. Some were L1 speakers of English; others were second language speakers. They broadly fell into Harwood et al.'s (2009) three categories of proofreader: professionals, who proofread as their main job; freelancers, who proofread less often and were mostly PhD students, proofreading for additional income, and volunteers, who proofread for free, mostly as a favour to friends. One of the proofreaders, Bernard, 6 was an inhouse proofreader, in that as an English language teacher, around a third of his job was taken up by one-to-one proofreading tutorials. In this article, I focus on a smaller subset of nine proofreaders for whom the theme of ethics was particularly significant as evidenced by their talk aloud/interview data, and/or their actual proofreading of the essay. Information about them can be found in Table 1, and it can be seen that all three of Harwood et al.'s categories are represented here.

Analysis of Proofreaders' Interventions
Proofreaders' interventions (i.e., their corrections, comments, and mark-ups/annotations when proofreading the essay) were systematically counted and classified using a modified version of Willey and Tanimoto's (2012) taxonomy, originally designed to analyse proofreading of L2 texts for publication. The taxonomy and quantitative results are in Harwood (2018). The taxonomy comprised eight categories, identifying the proofreader's Additions (adding words to the writer's text), Deletions (removing words), Substitutions (replacing shorter stretches of the writer's text), Reordering (repositioning the writer's words/sentences), Rewriting (replacing longer parts of the writer's text), Recombining (combining of one or more sentences, or division of longer sentences into shorter ones), Mechanical Alterations (changes to punctuation, spelling, and formatting), and Consultation/Teaching Points (where proofreaders address questions, comments, or suggestions to the writer).

Analysis of Interview Data
After summarizing the content of several interviews, a draft codebook was piloted, modified, and the final version consisting of 22 different codes was subjected to an intra-rater reliability test a fortnight after the first round of coding, resulting in an agreement rate of 75%. I felt this to be satisfactory, given the relatively large numbers of codes and the complexity of the scheme. NVivo was used throughout the procedure of coding and analysis, and the codebook included an ETHICS code, the focus of this article. In order to proofread ethically, informants spoke of how they adopted two main macrostrategies: (i) proofreading selectively; and (ii) declining to intervene. I present and exemplify findings related to these two strategies below, using proofreaders' interventions, talk aloud, and interview data.

Proofreading Selectively
In a number of situations, informants proofread less comprehensively than they could have done, for ethical reasons.

Proofreading Selectively to Develop Rather than Spoon-Feed Students
Rather than making a text 'perfect', Helena tended to ask writers many questions in an attempt to make her interventions formative, to enable students to 'try to develop' as a result of reflecting upon these questions, as she explained below: You get paid, you make [the writer's text] perfect, and they get 75%, which is an excellent mark. Then what would they think? That they are great.
[…] That they can easily stay in…academia.
[…] We should have no place for such people in the academia, and they have to realize it. So you either learn what your limitations are-you find out and you try to develop-or out of this place, please. 7 In Helena's view, those students who are not prepared to invest the necessary time and effort in self-study to enhance their academic literacy skills do not belong in the academy, and should not be spoon-fed by a proofreader.
An example of how Helena proofread selectively, questioning the writer and placing the onus on them to improve their text, centres on the following passage, where the writer in her essay is reviewing an experimental study and talking about its research design: Participants finished writings in four separated weeks using the provided linguistic forms. This experiment shows several result.
Helena said in her talk aloud that the meaning of 'four separated weeks' is unclear, and that she would need to read the source text the writer is reviewing to understand the intended message. When I asked for more detail about this episode at interview, I was surprised to learn that Helena wouldn't reveal she had read the source text to the writer; but she added that her covert reading would enable her to judge whether the student's rewrite, drafted in response to Helena's questions, was correct. Writers need to 'help themselves', and Helena would judge it to be 'immoral' to amend the text based on her knowledge of the source: Helena: I will pretend that I haven't read it. Int: …you won't correct it? Helena: No, no. No, no, no, no. They have to clarify it themselves. Int: Right…. So what you're doing, I think, is that you're reading it so that you understand it so that if the student goes away and rewrites it, you'll understand whether the rewrite is correct or not? Helena: Yes.
[…] That's why I say I will help them to the extent that they are willing to help themselves. I'm not going to do the job for them. For the reason that it is completely immoral for me to do that….
Thus the writer's role is to take responsibility for their writing where they are able. Helena gave two additional examples of what she meant and where she would decline to intervene: (i) where the writer has failed to follow a consistent referencing formatting style; and (ii) where the writer has omitted a date when citing in the main body of the text. Here is her commentary on (i); but in both cases, she stresses the necessity for writers 'to learn to work themselves': [Looking at the writer's list of references and their inconsistent formatting] It is pretty obvious that what the writer did is they copied those different [references] from where they saw them, so different authors follow [different referencing] systems or have their own way of doing things, and…what is the difficult thing to be consistent here? […] Even if it was a paid job I would…write a comment: 'Decide which system you want to follow and be consistent'. I might send a couple of links about this citation system. […] So no, I would not do the things that are so easy for you.
In sum, Helena wished to avoid the type of proofreading that meant writers were not obliged to develop their own academic literacy.
Sally also required writers to do some work themselves for ethical reasons. One of the resources given pre-task to proofreaders was the department's handbook containing details of the preferred referencing style to consult if they wished. But rather than consulting the handbook and adjusting the writer's in-text citations accordingly, Sally nudged the writer to do so, 'Cos I figure that's something they can figure out on their own and ought to know'. Sally took a similar view when she came to the writer's reference list, saying in her talk aloud: I'm not going to check the references. It's not my responsibility. Although by her own admission Sally nevertheless 'had a quick look through' the writer's reference list, she explained that amending it was the writer's role: You can copy the references from anywhere, you can look them up in a book, you can look them up on the Internet and make sure you've got them right. So I figure that's not necessarily something that I should be expected to look at in detail.
[…] For the rest of it, [being an L1 speaker is] an advantage I have over this person in that I know stuff that they don't. I can help them by providing that. But for this, I can look it up or they can look it up; well, they can look it up-it's their piece of work. […] I know that all the information is in the [handbook] and I feel that it's every student's duty or role to learn how to do referencing properly, and I don't think it would be helpful if I took over that….
Proofreading Selectively to Provide a True Picture of the Writer's Ability When deciding how comprehensively to correct, Bernard considered the writer's year of study. During his talk aloud practice and also in the main task, when debating whether to correct an error, Bernard said: It depends on the level the student is at; if it's a level 1 student I might let them get away with it. I will highlight it. Here are the two passages in question; the writer's meaning can be retrieved, but rewriting could improve the text: At interview, Bernard explained he intervened less in the case of first year undergraduates (level 1 students) compared to master's students, ensuring both the writer and their marker recognized students' true proficiency levels. This helped both parties understand how much work students needed to invest in language improvement and avoided the proofreader masking writers' actual (modest) abilities: I want it to be clear in their minds and in their lecturers' minds what their level of English is. And if I completely polish a piece of writing at a very early stage and they hand it in, both their lecturer and they think that 'Well, that's fine: they don't need to go to English classes.' So I like to leave mistakes in as long as they don't inhibit comprehension….
Recall that Bernard is not only responsible for offering students in-house proofreading tutorials; he also teaches pre-and in-sessional English classes, linking these classes to the proofreading tutorials inasmuch as he will cover grammar problems in class which he notices during proofreading. However, attending English classes is voluntary for students-unless their lecturers push them to go. Bernard therefore alerted both parties to the need to attend classes when he left errors uncorrected.

Avoiding Appropriation of the Writer's Text
To prevent appropriation, proofreaders tried to resist the temptation to improve the text's arguments, ensuring they retained the writer's voice. As she began reading the writer's introduction, Sheila encountered a sentence she found problematic. In her talk aloud, Sheila said: I can't rewrite the whole sentence, but highlighted it in yellow, and nevertheless provided a substantial rewrite: Truscott's(2008)  At interview, Sheila explained the question marks and highlighting signalled the provisional nature of her rewrite:

Writer's text Sheila's version
…and I make sure that there's question marks there, so that's like an error code that just says 'This is how you could rewrite this previous sentence…,' '…you have to make a decision, check that the meaning is unchanged…and preferably use your brains and rewrite it yourself along these lines.' The decision regarding how the final text should look was therefore in the hands of the writer-although Sheila acknowledges there are ethical issues around even providing a rewrite for the writer to rework herself: I always tell them that 'If I do that [i.e., suggest a rewrite], technically you read it and make a decision as to whether you want to put it that way, but it really should be something you've written.' But I know she can't or he can't write it this way because they wouldn't construct a sentence this way. So there's a fine line between rewriting the whole essay and making suggestions for how sentences should sound and I would hope that they would do the second.
Here Sheila is conscious of the possibility of introducing rewrites into the writer's text to a standard the writer would be unable to produce independently.
Eleanor also spoke of the importance of resisting the temptation to rewrite for elegance, so as to ensure the retention of writers' ownership of the text. She justified this by conceptualizing her role as a proofreader rather than editor; and for reasons of time: …even if I think that there's an even better word to use, as long as the meaning is clear…and they're not being overly repetitive, then I would just leave it in their own words as much as possible.
[…] It would take a lot longer if I was going to put it all in my own words, and also it would then take away the fact that this is the student's work….
Other examples of unacceptable practice for Eleanor were 'structural changing' and enhancing the quality of writers' arguments and content. She resisted giving writers 'information that they haven't already thought of or researched themselves', despite on occasion proofreading texts from a field with which she is familiar when she could have made content-level suggestions: So in the [name of discipline] thesis that I've been proofreading, there are things in there that I know that the student hasn't mentioned, but I can't mention them. So that just has to be left.

Declining to Intervene
I now turn to why proofreaders decline to work with certain texts and certain writers.

Declining Unscrupulous and Inappropriate Requests by Writers
Helena claimed many L2 students who request proofreading have inadequate language skills for study purposes, but 'don't have the time or the interest' to enhance them, instead preferring to pay someone to bring their English up to standard, to strengthen their content and argumentation, and to boost their mark. Meeting these requests would therefore mean proofreading unethically. Helena begins by talking about her earliest experiences of being approached to act as this sort of 'proofreader': …those people…were international students…, whose English was not good, and they didn't have the time or the interest to develop it; they just wanted to get their Master's and leave this country. …maybe they didn't work hard on their topic, or they did things in a hurry. And everything was so messy […] I realized that they don't expect me to correct the grammar, the punctuation, or maybe some vocabulary, but they would expect me to intervene and contribute to that, and… Int: You mean in the level of content? Helena: Yes. Yes. Int: You mean like add things? Helena: Yes. And their argument was 'I will pay you.' […] and they would expect me to make it look better: to rewrite it.
Helena's stock response now is to decline these requests: So they would assume that…I will… transform that [messy] piece of writing to a nice essay with which they will pass and have a nice mark. No. I'm not willing to do that. […] So the only case that I would take a proofreading job would be if I saw the writing and all that they would need is help with punctuation, maybe with cohesive devices…, with a bit of grammar, a bit of vocabulary: I can do these things, and I really enjoy it, and I can also have peace of mind that I didn't do anything immoral.
The proofreading task text is typical of the poor quality texts she now declines (unless she helps a friend as a favour). She believed the essay may be graded as a fail, and there was far too much work needed to proofread it ethically: So even if you polish this with the right vocabulary and punctuation, how much will it gain? Why shall I accept money for that? […] And I don't think that the person who will ask me to proofread this particular piece…would mean just put the capital letters and the inverted commas right.
Helena shared a personal experience to support her view that writers expect proofreaders to boost their marks. Helena proofread her friend's essay, which was awarded a mediocre mark of 55. The writer had expected a much higher mark, despite Helena's advice, which was disregarded, that more work was needed before submission. As Helena argued, then, her friend apparently believed that the 'proofreading' would transform the text with little or no reciprocal effort on the writer's part: The introduction, the structure, was not okay; the main body, the arguments, were not okay…; and the conclusion was not even a conclusion really. So I told her…she has to work a little bit more…, and she told me 'I don't have the time…, please do your best.' So I…did my best to correct the English and rephrase what she really wanted to say, because her English was a bit poor.
[…] And then when she got the mark she said, '…I was given only 55 for that.' And I said, 'OK, so what? 55's not bad. It's too much for that.' She said 'But you corrected it.' I said, 'Yes, I corrected it, but I didn't write it from scratch. I told you that there are…other things that you have to correct. It was not only the grammar and the vocabulary'.
[…] She said, 'I thought I would get more because you corrected it.' Linda had also received and declined unethical requests. One writer asked her to 'look things up, research things for him in the library and I said 'I can't do that cos I'll be doing your work…I'm not here to actually research your [writing]'.' The same writer subsequently asked her to paraphrase someone else's writing 'on a similar topic to his', passing the work off as his own. Again Linda declined. Indeed, Linda described how some students were 'willing to pay anything' for ghostwriting: And then I have another guy that near enough wants me to do the essay for him. I've had people like, 'I'm running out of time, can you do it? Can you finish the essay off?' […] And I have to be quite… determined to tell them that that's not what I do….
Like Linda, Sheila also had what she called 'Please write this for me' stories. As we saw above, Sheila may ask for rewrites of parts of a text where the writer's intended meaning is unclear, requiring the student to supply the rewrite themselves for ethical reasons. However, sometimes writers tried to pressurize her to supply the rewrite herself.
The proofreaders' two macrostrategies, proofreading selectively and declining to intervene, are described and evidenced above. I now present two other important themes relating to ethical issues which emerged from the data. The first of these relates to how, despite their attempts to proofread ethically, for some proofreaders there were still dilemmas and uncertainties which loomed large during their work. The second concerns the degree to which informants were willing to intervene: some proofreaders intervened at the levels of argumentation, structural organization, and content in ways which went far beyond traditional conceptualizations of proofreading-and which could be construed as unethical.

Proofreaders' Disquiet and Self-Doubt
Having made a number of changes, Fiona wondered in her talk aloud whether I am going beyond my remit here. She was referring to places where she made substantial changes to the text, as in the following:

At interview, Fiona explained her 'beyond my remit' remark was
…because I decided to write...whole chunks myself…. I always imagined that proofreading would be putting like a semi-colon or making sure that the apostrophes are in the right place, and it's turned out some of these students who are non-native English speakers, I'm actually writing the damn thing…, but I sometimes think I'm stepping over the mark.
As a retired lecturer, Fiona took a dim view of the university's decision to admit lowproficiency doctoral students whose essays she claimed she would fail even at undergraduate level; but by proofreading for these students, 'I'm actually part of the whole thing, I'm condoning it'. Fiona also argued that in her 'proofreading', she was having to teach fourthyear PhD students academic writing conventions that their supervisors should have covered, such as the need to cite authoritative sources. Although Helena mostly expressed her views on unethical proofreading forcefully and confidently, self-doubt wasn't entirely absent from her reflections. We saw earlier how, where the writer's intended meaning was unclear, Helena would intervene by asking questions in order to have the writer clarify their intended message ('Did you mean this?' 'What are you trying to say?'). Helena would then respond to the revisions as long as the writer kept the dialogue going. However, in the extract below, Helena wondered whether she was inappropriately performing the role of the lecturer/supervisor: Helena: I'm happy to work with them as long as they're happy to answer my questions and… Int: So as long as they keep working on it, you keep working on it? Helena: Yes. And then I sometimes question myself and I say 'But you're not their supervisor…, it's not your job…'.
Sally also expressed uncertainty and disquiet about the boundary between acceptable and unacceptable proofreading. She reported finding the proofreading task 'quite difficult and quite frustrating'; the text was 'atrocious' and she felt she was 'remoulding' rather than simply 'polishing' it-polishing being what she does with the higher quality texts she normally proofreads. Here though, she experienced ethical dilemmas about how far to intrude: …how much is it my responsibility to simply correct language and how much should I get into the structure of it…? And is your argument secure and sound? And all those broader teaching issues which I try not get involved in if I'm just reading through it to check the language.
An example of Sally's uncertainty can be seen in her question to the writer in response to the excerpt below:

how do you know?
In her talk aloud, Sally remarked when trying to decide how to intervene here: This is where I find it difficult because I'm not sure whether I'm just supposed to be correcting the English. At interview, she reflected on this episode and on her conceptualization of proofreading boundaries further: …to what extent should I get caught up in questions of structure or content or 'Have you answered the question?', 'Is this relevant?' and that kind of thing? Or whether I should just ignore all those questions and simply take it at face value: this is what they've written, just tidy it up and make it clearer? But in an example like this where it's no hassle or effort for me to just raise the question then I would usually do so cos I think it's important.
In the example above then, Sally did indeed comment on the writer's (unconvincing, incomplete) argumentation ('how do you know?'), albeit as she questioned whether she should be intervening at this level at all. However, at other points in her proofreading, Sally declined to become involved for two reasons: (i) ethics; and (ii) time. At interview, I asked for reflections on another of her talk aloud comments which touched on the question of ethics: I don't feel I'm doing a terribly good job because I don't really understand what they're trying to say. But I don't think what they're trying to say is terribly good. But I suppose that's not my responsibility: …if I had a thousand years to do it all…it would be nice to do it really thoroughly…and really check, do it in teacher mode, Have they answered the question? Have they structured it sensibly? Does this relate back to things that they've raised earlier on?, etc, but I have to hold myself back from getting into all of that cos I think, A) it's not what I'm supposed to be doing and B) it would be far too time consuming…. So trying to…just accept, well this is what they've tried to say, so this is what I'll help them to say, even if I think that perhaps it's not very clear or doesn't really do justice to the question or the arguments that they've presented or whatever.
[…] It's somebody else's work, it's what they've wanted to say and try and keep it as their piece of work…. But just improving the surface effect.
Going Beyond 'Proofreading' in Ethically Questionable Ways Some proofreaders' practices go far beyond traditional understandings of proofreading and can be seen as ethically questionable. The proofreader most notable for her argumentation and content interventions was Ana. Consider her comments on the following paragraph:

Add one line or two defining what the error code is.
Ana explained at interview that the writer's argument lacks clarity and explicitness here, and she argued that it was acceptable for proofreaders from the same discipline as the writer to help with the content/ideas, as well as the language of the text.
We saw earlier how Sheila believed the onus was on writers to rewrite themselves; nonetheless, she also sometimes commented on the persuasiveness of an argument. And so despite not supplying a rewrite, these were still interventions at the level of content. One such example can be seen below:

(Perhaps reconsider this example.)
As can be seen, Sheila highlighted the sentence and asked the writer to think again about what they had written. She explained her intervention at interview: I wouldn't use this example because she said the error codes can be really complex, but this is hardly complex. G for grammar, SS for sentence structure, and SP for incorrect spelling. That's just three things, maybe she means texts where there are 12 different codes…. That's complex but this is not, so maybe she should just delete that.
Here is another example of where Sheila focused on the persuasiveness and relevance of the writer's argument, questioning the level of detail the writer provided about the experiment being reviewed: Linda described a writer's request which she conceded 'is not really proofreading'-but which she nevertheless agreed to take on. We saw earlier how Linda had been approached by a writer who wanted her to ghostwrite and was refused; but he then asked Linda to 'move things around' and this request to restructure the text was accepted:

Writer's text Sheila's comments
…so it's more a question of editing rather than proofreading, so I suppose that's different and I do that. That's not strictly proofreading, there's more depth to it and it's moving chunks but he wants me to do that, so I do that.
A final example of ethically questionable interventions emerged during Moira's interview. Moira explained how she would ask the writer to send her the source texts and to underline the parts of these texts s/he had used in the essay when submitting work to her. I wondered what the significance of this request was and asked for further details: Int: But to do what? Why do you want to see what the source is? Moira: Well, to make sure it's fine in the context of how they've used the citation in it. Or they're not making sweeping statements, and we know that it shouldn't happen but it does, people just plagiarise and they just take bits they forgot...are not sourced properly and so I always check.
Moira would therefore check the writer's summaries and reviews of the sources cited for veracity and for plagiarism-again, clearly going beyond traditional conceptualizations of grammar and syntax proofreading.

Discussion and Implications for the Implementation of Ethical Proofreading Practices
My findings relating to the ethical dimension of proofreading lend weight to some of the pro-and anti-proofreading arguments outlined earlier by McNally and Kooyman (2017). One danger is proofreaders performing inappropriate interventions. Although the more extreme requests (e.g., ghostwriting) were declined, several of my informants were prepared to some degree to intervene at the levels of structure, argumentation, and content: Linda took on structural editing, Moira checked writers' sources to ensure the writer had accurately represented them, and Sheila and Ana identified rhetorical weaknesses in the sample text and signalled ways for the writer to buttress her claims. In these cases, then, if it is too much to say that proofreaders are appropriating writers' texts, they are at least providing them with suggestions and sometimes even rewritten sentences to strengthen their texts in ways far beyond correcting grammar.
Turning to McNally and Kooyman's pro-proofreading arguments, there was evidence proofreaders attempted to make their interventions formative, and that some informants proofread very selectively so that the writer's true level of linguistic competence was apparent to lecturers (and students). The results also confirm the importance and prevalence of the ethics theme suggested by Harwood et al. (2010); but whereas Harwood et al.'s study was wholly interview-based, the present research presents more specific, concrete evidence in the form of textual interventions and talk aloud data of how ethical questions play out during proofreading, in addition to reported behaviour. It would however be misleading to claim that the ETHICS code was evenly distributed across all informants equally. Just as in Harwood et al.'s (2010) study, some informants said little about ethics in general or experiencing ethical uncertainties in particular, whereas others, like Helena and Sally, had much to say on these matters. Still others like Ana said little, but it was clear from their interventions at the level of structure, argumentation, and content that ethical questions arise in relation to their conceptualization and execution of prooofreading, in that they go far beyond the boundaries of traditional proofreading as defined by the Society for Editors and Proofreaders (2005).
I close by discussing the implications of the findings of my study, and argue that three principal messages emerge: (i) the case for regulation of proofreading; (ii) the need for effective dissemination of institutional guidelines on proofreading; and (iii) the desirability of the promotion of in-house, formative proofreading.

Formulating and Disseminating Proofreading Regulations
When Harwood et al. (2009) conducted their research at the University of Essex, they found that the proofreading of student writing was totally unregulated, meaning that institutional guidelines on permissible forms of proofreading were entirely lacking. Happily, the situation has now changed-not only at Essex but at many UK universities, and the drawing up of institutional regulations and guidelines for proofreaders, writers, and lecturers seems like an important first step on the road to confronting ethical issues associated with proofreading.
However, just because proofreading guidelines are in place, this does not mean proofreaders follow them. My research was conducted in an institution which had proofreading regulations in place (specifying that interventions should be relatively light-touch, prohibiting corrections or comments at the level of argumentation, structure, and content). However, the majority of proofreaders were apparently unaware of these regulations (and as discussed above, some clearly breached them). The exceptions were Eleanor, who knew the university's regulations didn't permit her to 'edit'; Fiona, aware there were regulations, but unsure whether she was violating them; and Bernard, who proofread in an official capacity as part of his role as an English language tutor. Whatever policies are specified, then, wide dissemination is needed so what is permissible is clear to all parties-lecturers, students, and proofreaders themselves.

The Nurturing of In-House, Formative Proofreading
Given that allowing proofreading to be 'outsourced' to freelance proofreaders may mean that these proofreaders are unaware of university policy, there is much to be said for keeping proofreading in-house as far as possible. This would make regulation and surveillance easier; compulsory standardization and training and development sessions could be held during which proofreaders would all respond to the same essay until a standard was established and boundaries were drawn up beyond which help could not be given, ensuring much more consistent treatment from proofreader to proofreader than I saw in the present study. Furthermore, as part of a standardized proofreading policy, universities could implement more ethical, developmental forms of proofreading operating in a similar manner to Bernard's approach, with the aim of teaching students about writing and language rather than merely fixing their texts.

Future Research: Establishing a Consensus for Ethical Proofreading Policies
Whatever kind of ethical proofreading policy is proposed, though, consultation with lecturers is required to ensure their support. McNally and Kooyman's (2017) recent study shows how this can be done; questionnaires were used to solicit the views of content lecturers, English language tutors, and students on various types of proofreading interventions. Encouragingly, there was a considerable degree of consensus regarding acceptable and unacceptable practices between the parties. However, there was less agreement regarding more substantial types of proofreading. One sample intervention supplied the writer with text rather than merely correcting or commenting upon what the student had written: Your introduction should finish with an outline, e.g. BThis analysis will look at the air temperature and extreme weather events in more detail, and then examine how these determinants…^.
The majority (62.7%) of students surveyed felt this intervention was 'just right'; but 73.4% of lecturers and tutors felt it went too far. Such an intervention is similar to some of Ana's comments which suggested changes to structure, argumentation, and content. Indeed, given this finding, one may imagine lecturers would feel Ana, Sheila, and Moira's interventions all go too far and that ethical proofreading should be defined more narrowly.
McNally and Kooyman's sample was modest (30 staff and 59 students), mostly from a small Australian institution with a population of predominantly non-traditional L1 rather than the L2 students focused on in my research. Furthermore, relatively few examples of proofreading interventions were shown to respondents. Lastly, because McNally and Kooyman relied on closed-format questionnaires to gather their data, in-depth responses are lacking. Future work can usefully build upon their study to provide us with a greater understanding of stakeholders' views on ethical proofreading. This work could combine quantitative and qualitative methods, using more detailed questionnaires containing a greater number and range of example interventions in the style of McNally and Kooyman's instrument, as well as semistructured interviews, enabling respondents to explain the reasons for their beliefs about ethical interventions. A focused study of this type could also design into the instruments a more detailed examination of the types of ethical theories underpinning stakeholders' beliefs about proofreading, focusing particularly on consequentialist and deontological perspectives (for a useful overview of these and other ethical theories, see Kaptein and Wempe 2011).
It will not have escaped readers' notice that there is plenty of evidence of consequentialist thinking in discussions of appropriate proofreading put forward in the literature I discussed earlier as well as in my own data. Hence McNally and Kooyman (2017), Bernard, Eleanor, and Helena worry that proofreading too assiduously will result in lecturers gaining an overly favourable impression of the student writer's ability and the award of a grade the writer doesn't merit. More consequentialist reasoning can be seen behind Bernard's decision to only correct selectively; he calculates that if the student (and the student's lecturer) is not alerted to the student's language problems, s/he may choose not to attend Bernard's optional in-sessional English classes. And it is emphasized that appropriate proofreading should seek to discourage unreflective responses on the part of the writer: if, for instance, more selective, indirect, or interactive forms of intervention are used, students will be obliged to take ownership of the revisions with a more educative outcome anticipated than if the proofreader adopts a 'fix-it' role which enables the writer to simply adopt the role of a passive onlooker. Indeed, the call for firm regulation of proofreading found in McNally & Kooyman with which I am in agreement is also consequentialist, in that it is designed to stop one student receiving a much more/less substantial form of proofreading than another writer using a different proofreader, who may be awarded a more/less generous grade as the result of the proofreader's work. It is likely, then, that much thinking about the ethics of proofreading is consequentialist. Yet when another proofreader, Sally, speaks of student writers' 'duty' to learn to reference properly and when Fiona speaks of her proofreader 'remit', such thinking carries more of a deontological flavour. And perhaps one could find some lecturers adopting deontological reasoning too. For instance, they may be of the view that all work should be the writer's own, given that university essays and assignments are commonly set as sole-authored projects and that collaboration and collusion during composition are often expressly forbidden. Another example of deontological thinking would be where lecturers feel that if students have met the university's admissions criteria, this signals their ability to cope with their writing unaided.
Then of course there are the ethical dilemmas which are experienced by proofreaders like Fiona in my data which move between consequentialist and deontological reasoning, on the one hand referencing the likely deleterious consequences of proofreading student writing in a particular manner-or not proofreading it at all-before submission, and on the other referencing the writer's and the proofreader's duties and responsibilities which are seen as self-evident or non-negotiable by the informant. Presumably other stakeholders (content lecturers, English language tutors, and student writers) could experience similar dilemmas when reflecting on their views of proofreading. Questions which duly exemplify consequentialist and deontological ways of thinking about proofreading could be added to questionnaire/interview instruments and enhance our understanding of the reasoning of all parties affected by proofreading. Alternatively or in addition, proofreading scenarios which describe ethical dilemmas typifying consequentialist-deontological tensions could be discussed. Each approach will enable a more focused, systematic, and in-depth comparison of ethical positions.

Introduction
During the recent decades, the effort of error correction and feedback become a more and more controversial issue. Truscott's (2008) mentioned that error correction, especially grammatical error correction has no or even harmful effort on improving accuracy in Second Language students' writing, because this kind of error correction could reduce teachers and students time and energy on more important and significant aspects, such as students' thoughts and structure.
The purpose of this essay is to discuss the different experts opinions of the effort in error correction in order to explore which one of these arguments, based on the experiments these experts employed, maybe considered more reasonable and sensible for developing accuracy in Second Language students writing.
Firstly, a brief introduction will be given to show the definition of error and different types of error correction feedbacks. Secondly, Truscott's opinion of error correction will be described with his supportive experiment. Following with some other experts arguments with their experiments to support their argument such as Ferris DR & Hedgcock JS, Bitchener, J. et al., Chandler J, Lalande.JF. After that, personal teaching context will be given to substantiate personal arguments.

Errors and Error Feedbacks
Ferris DR & Hedgcock JS (2005) define that 'Errors consist of morphological, syntactic, and lexical deviations from the grammatical rules of a language.'. Usually, Second Language writers have trouble with 'verb inflection errors', 'English determiner system' and 'word order', such as verb tense, aspect, voice; subject-verb agreement; and active or passive constructions, etc.
There are three types of feedback for teachers to correct errors. Firstly, selective error feedback which is focus on Second Language students' most serious and frequent patterns of errors and comprehensive error feedback which teachers need more time and consideration to concentrate on. Secondly, error feedback on larger categories and error feedback on smaller categories. Error codes are used in both larger categories and smaller categories.
However, teachers and students need time to familiar with these various and complex error codes. For example, 'G' means grammar error, 'SS' means sentence structure, and 'SP' means spelling, etc. Thirdly, direct feedback which teachers correct errors directly on the original draft and indirect feedback which students are required to self-correct with or without underlined errors. Truscott's (2008) indicates that error correction, especially grammatical error correction (as one of the most controversial issue in error correction) has little or no efficiency on developing accuracy in Second Language writing. Because error correction could enforce students and teachers to focus on and reduce their energy and attention from other aspect in writing, such as students' thought and composition structure.

Truscott's Opinion of Error Correction
In order to support Truscott's view, he design an experiment (2008): forty-seven students were divided into two groups-experimental group and control group-to be enrolled in three sections of a writing seminar. After the first article, experimental group received their article with errors underlined and need to revise their article, but controlled group received no-marked draft. It is easily to see that experimental group performed better than control group on revisions and error feedback made a positive effect on students' rewrite.
One week latter, students had a new article. Compared error rate of the second article with the first one, both group received the equal results, that is to say, there is no positive effort on experimental group second article although they have their first article revised and rewritten. There are some limitations in this experiment: firstly, students could gain knowledge in a short time; secondly, the Truscott's experiment only focus on one type of feedback. At last, Truscott draw a conclusion that there are no evidence to show the effectiveness and efficiency on improving students' writing ability by correcting errors.

Lalande's View on Reducing Students' Errors
Lalande (1982) contends that there are four strategies which could have an effective influence on developing students' writing skill. For instance, 'comprehensive error correction' with which students could fully improve their skills (although this kind of correction could take more time and energy from students); 'systematic marking of composition' which would effective in reducing errors of students' compositions; 'guided-learning and problem-solving' which could encourage students in Second Language writing abilities; 'instructional feedback' on which error codes were used to show the location and nature of errors.
Lalande employed an experiment to measure the effectiveness and efficacy of these four strategies on 'grammatical and orthographic correctness' of Second Language writing. Four classes were divided into two groups-experimental group and control group. Lalande collected date before the experiment to ensure that there is no important and considerable differences between experimental group and control group. The feedback was also be strictly controlled and no detail or information should be involved in the feedback. A large number of short articles had been read by students in control group and teacher of control group give comprehensive corrections on students' article and demanded for 'incorporating' by same aspects. The error code and 'error awareness sheet' were used in experimental group students' writing for them to realize the nature of error and to understand deeply immediately before the next article. As a result, students in experimental group developed their 'grammatical and orthographic' abilities much more than students in control group. There are some limitation in this experiment: firstly, subjects may not been chosen that random; secondly, more students should be enrolled in the experiment. Bitchener (2005) states that incorporate different sort of correction feedback such as oral feedback and written feedback could improve students' writing abilities, especially linguistic error corrections, not only improve in the original rewrite essay but also another new essay as well.

Bitchener's View on Various Kinds of Correction Feedback
To support Bitchener's issue, he designed an experiment and 53 new students were acted as participants into this experiment. The participants has been divided into three units by different educational time. The students of the first unit who gained the longest educational hour could receive direct written correction feedback and a short time students-teacher tutorial which students and teacher could discuss the unsure issues and example of the essay and then teacher would give students extra or further examples or textbook questions with the same type of errors as exercises. The students from the second unit who obtained the moderate educational hour could receive direct written correction feedback but no tutorial combined. The students from the third unit who had the limited educational hour chould receive feedback about their 'quality and organisation of content'. While during classes, teachers could discuss some 'form of instuction' as part of courses.
Participants finished writings in four separated weeks using the provided linguistic forms. This experiment shows several result. Firstly, it could be easily commend that the improvement of individual feedback are different due to the different time of writing. Secondly, the feedback gain a small effect according to the separation of targeted linguistic forms. Thirdly, the indirect feedback has more positive effort than direct feedback when students improve accuracy by write another essay. Fourthly, direct oral feedback connected with direct written feedback showed the significant influence than any other type of feedback. The last but not least, 'rulegoverned linuistic features' are easily improved by oral feedback connected with written feedback. However, further research would be needed to investigate long-term accuracy.

Chandler's Research About Error Correction and Accuracy
Chandler (2003) noted that students improve accuracy when they are recommended to selfcorrect and self-edit 'grammatical and lexical errors' after receiving teachers feedback. And she also wanted to find out the result if students correct latter after receiving the teacher's error correction feedbacks.
Chandler employed a study to find out the relationship between error correction and accuracy. In her experiment, the two classes students were asked to write about five types of essays. And the only difference between experimental group and control group is that experimental group were asked to to self-correct their errors which teacher had underlined before submitting next essay. However, control group selfcorrect all their errors at the end of semester. Ten weeks latter, she found out that both the experimental group and control group improved in fluency over the term. However the control group which did not correct their errors between each essay did not improve their accuracy while the experimental group has a positive effort on accuracy after self-correct between each essay. It is also the fact that if students did not self-edit or self-correct their errors after receiving feedback from teachers the result is equal to receiving no feedback. There is no improvement between nonfeedback and non-correction.
The Agree and Disagree Argument Between Truscott and Ferris Truscott (1996) state 'grammar correction' as 'correction of grammatical errors for the purpose of improving a student's ability to write accurately.....correction comes in many different forms, but for present purposes such distinctions have little significance.' However, Ferris disagree with this argument. She (1999) mentions that error correction does have positive effort on error correction according to many research evidence. Then, Ferris pointed out three main mistakes of Truscott's review:(1) The themes could not to be contrast because they are in different studies and based on different experiments. (2) The investigation and strategies changes thought the different research. (3) By ignore the effort of research, Truscott's passive evidence could not controvert his statement. She also remarks that problems which teachers may not recognize an error, or teacher could not explain the error,or even teacher explain the error but students may not understand that error could be conquered. At last, Ferris claimed that teachers and students should not avoid error correction only because students do not develop their self-correct or the shortage of teachers error correction feedback. Ferris (1996) agree with Truscott (1996) that 'syntactic, morphological, and lexical knowledge' are seperated by different categories, so one structure of error correction could not suitable for all of these three types. And she also suggested that the significant and necessary of correction, practicing on recognize and correct the 'frequent and serious errors', clarification the rules of error 'patterns' could improve teaching self-correction. At the end of her argument with Truscott, she appeal for further research.
Ferris's Suggestion on how to Gain Accuracy in Grammar Error Correction Ferris (2004) suggested honestly to both teachers and students how to treat error in students' writing. First, the attendance of lesson and reading book-based materials and webbased materials, practicing the recognize errors from students' exam paper and course works, familiar with grammar knowledge and corrective abilities could encourage teachers ready for correct students' error effectively and efficiently. Second, focusing on students' desire and educational background information when teachers create error correction feedback because error correction is not the unique aspect of students' writing. Third, 'linguistic accuracy and editing skills' could be gain not only by error feedback but also by social activities.
There are six suggestions as follow. (1) Error feedback is one of the essential aspects in students' writing, so teachers need much more motivation on devising courses and take error correction seriously. (2) Indirect error correction feedback could encourage students' automaticity in self-correction. (3) Some error may be unsuitable for students' self-correction, such as 'lexical errors, complex and global problems with sentence structure'. (4) Revision is considerably necessary for students to find out their weakness and drawbacks. (5)'Grammar instruction' could be easily reduced in accuracy with other sources of error treatment. (6)'Error chart' could enhance students attention of drawbacks and development of writing.

Personal Teaching Context in Error Correction
During the pre-sessional period, error correction became a novel aspect with enormous influence in academic writing because multiple choice, spot dictation and comment has become three main aspects in China examination, however, assignment do not included in education system.
The error codes were hard to familiar at the beginning, so the checklist of error codes information is extremely suitable for a beginner. Error correction, such as 'grammatical and orthographic correctness' were not that important and significant in pre-sessional period, error codes were usually employed in the essay followed by underlined errors which students need to self-correct. Tutors were usually focus on the structure and organization of the essay. And detailed feedback was divided into several aspects, for instance, overall issue shows the improvement for the former draft; introduction focus on the proficiency of introduction which is useful for readers have an overview of essay and understand the importance of the essay issues; 'academic line of enquiry' shows the abilities of using relevant according to the topic of the essay; 'reporting of ideas from source texts' is about the student's personal ability to summarize and paraphrase; language and style states the development and improvement of the syntactic structures and academic vocabulary; conclusion focus on the abilities of summarize and related to the essay topic.

Conclusion
These decades, the argument of whether error correction could developing the accuracy of Second Language students writing becomes more and more crystallizing. The important and significant role which error correction plays changes the teaching strategies of English language.
The benefits and inadequacies of using error correction for students' writing has been discussed in this essay based on the arguments of different experts to show that error correction do has important and significant effort on efficiency and effectiveness of accuracy in Second Language student's writing. So Truscott's criticisms of error correction was not supported in this essay.
Firstly, 'error' was defined at the beginning of this essay and the different categories of error correction also be located. Secondly, Truscott's issue that error correction do not have positive effort on accuracy in Second Language students writing was stated and his experiment also be employed to support his argumentation. After that other experts opinions such as Ferris's, Bitchener's, Chandler's, Lalande's were supported with their experiments. Finally, personal teaching context was pointed out to emphasize that the important and significant role error correction plays in improving accuracy in Second Language student's writing.
To sum up, from the previous explanation of error correction followed by the discussing of several experts opinion, it is clearly noticeable that in developing students' writing, using error correction could enhance student's efficiency and effectiveness in accuracy. Here are several metaphors for describing the role of a proofreader. Can you comment on each of these, saying what you understand by each metaphor and how far you agree or disagree with each metaphor?
The proofreader is a CLEANER or TIDIER HELPER or MENTOR LEVELLER, lessening the disadvantages for non-native writers MEDIATOR, bridging the gap between student and lecturer TEACHER 14. Finally, is there anything else you'd like to say or comment on relating to proofreading?
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.