Sugaring the pill: praise and criticism in written feedback

Journal of Second Language Writing 10 (2001) 185 – 212 Sugaring the pill Praise and criticism in written feedback Fiona Hylanda, Ken Hylandb,* b a University of Hong Kong, Hong Kong, China English Department, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong, China Abstract This paper offers a detailed text analysis of the written feedback given by two teachers to ESL students over a complete proficiency course. We consider this feedback in terms of its functions as praise, criticism, and suggestions. Praise was the most frequently employed function in the feedback of these two teachers, but this was often used to soften criticisms and suggestions rather than simply responding to good work. Many of the criticisms and suggestions were also mitigated by the use of hedging devices, question forms, and personal attribution. We explore the motivations for these mitigations through teacher interviews and think-aloud protocols and examine cases where students failed to understand their teachers’ comments due to their indirectness. While recognising the importance of mitigation strategies as a means of minimising the force of criticisms and enhancing effective teacher – student relationships, we also point out that such indirectness carries the very real potential for incomprehension and miscommunication. D 2001 Elsevier Science Inc. All rights reserved. Keywords: Teacher feedback; Praise and criticism; Mitigation; Miscommunication Introduction Providing written feedback to students is one of the ESL writing teacher’s most important tasks, offering the kind of individualised attention that is otherwise rarely possible under normal classroom conditions. However, while * Corresponding author. Tel.: +852-2788-8873; fax: +852-2788-8894. E-mail address: [email protected] (K. Hyland). 1060-3743/01/$ – see front matter D 2001 Elsevier Science Inc. All rights reserved. PII: S 1 0 6 0 - 3 7 4 3 ( 0 1 ) 0 0 0 3 8 - 8 186 F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 generally acknowledged as pedagogically useful (e.g., Cohen & Cavalcanti, 1990; Hedgcock & Lefkowitz, 1994), the role of written feedback has largely been seen as informational, a means of channelling reactions and advice to facilitate improvements. Because of this, its important interpersonal aspects, the part it plays in expressing a teacher’s stance and beliefs about writing and in negotiating a relationship with learners, is often overlooked. As teachers, we are usually conscious of the potential feedback has for helping to create a supportive teaching environment. In addition, we are aware of the need for care when constructing our comments. We know that writing is very personal and that students’ motivation and self-confidence as writers may be damaged if they receive too much criticism (e.g., Connors & Lunsford, 1993). We may also believe that praising what a student does well is important, particularly for less able writers, and we may use praise to help reinforce appropriate language behaviours and foster students’ self-esteem. However, while teachers may recognise that the use of praise and criticism in feedback is important, the use of these features as central resources for negotiating judgements and evaluations of student writing and their contribution to the pedagogic role of feedback in EFL contexts has not been systematically studied. In this paper, we offer a detailed analysis of the ways praise and criticism are used in the written feedback given by two teachers to six ESL students in a university language enhancement course. Following Holmes’ (1988) characterisation of compliments, we view praise as an act which attributes credit to another for some characteristic, attribute, skill, etc., which is positively valued by the person giving feedback. It, therefore, suggests a more intense or detailed response than simple agreement. Criticism, on the other hand, we define as ‘‘an expression of dissatisfaction or negative comment’’ on a text (Hyland, 2000a, p. 44). This definition thus emphasises commentary which finds fault in aspects of a text, and we felt the need to distinguish this from a third category, suggestion, which we regard as coming from the more positive end of a continuum. Suggestions differ from criticisms in containing an explicit recommendation for remediation, a relatively clear and accomplishable action for improvement, which is sometimes referred to as ‘‘constructive criticism.’’ Our definitions thus draw heavily on common sense understandings of these terms, but also embody distinctions that were recognised by both student and teacher participants in the study. In what follows, we examine the forms and patterns of these acts and draw on classroom observations and interviews to uncover teacher and student perspectives on the ways they are used, before discussing some implications for teachers. Importantly, we see praise and criticism as not only a crucial feature of the teaching and learning context, but also as helping to constitute this context. The teachers in this study were well aware that the type of comments they gave acted to both directly contribute to learning and to create the interpersonal conditions in which learning might occur. F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 187 Significant prior research Most of the research on praise and criticism has occurred within a framework based on politeness, has examined speech, and has focused on complimenting behaviour (e.g., Herbert, 1990; Holmes, 1995). Our knowledge of how criticism is expressed is similarly very limited and largely restricted to (often elicited) conversational routines among intimates (e.g., Beebe & Takahashi, 1989). Politeness models have also been influential in writing in explaining how particular features are used to maintain rapport and mitigate criticism in student peer reviews (Johnson, 1992) and, more recently, in published academic book reviews (Hyland, 2000a). The use of praise and criticism in written feedback is, however, more complex than in book reviews. Teachers are usually not simply appraising writing, but are often hoping to use the opportunity for teaching and reinforcing writing behaviours. In fact, they may be fulfilling several different and possibly conflicting roles as they give feedback: sometimes acting as teacher, proofreader, facilitator, gatekeeper, evaluator, and reader at the same time (Leki, 1990; Reid, 1994). In addition, their personal knowledge of the writer is usually greater than it would be between a book reviewer and an author, and they probably have more interest in creating and maintaining a good face-to-face relationship with the student. In other words, teachers often have to weigh their choice of comments to accomplish a range of informational, pedagogic, and interpersonal goals simultaneously. Research on praise and criticism in feedback is fairly sparse. Several L1 studies suggest that teachers attend to error more than excellence and tend to focus their feedback on the negative aspects of the writing. Dragga (1986, cited in Daiker, 1989), for instance, analysed 40 student essays and found that 94% of comments focused on what students had done poorly or incorrectly. Experimental studies have often gone further to examine the different effects of focusing on positive and negative aspects of texts. Taylor and Hoedt (1966), for example, failed to find any difference in the quality of writing produced by students receiving either positive or negative feedback, although they did show that negative feedback had a detrimental effect on writer confidence and motivation. Gee (1972) also reported no significant differences in quality of writing, but more positive attitudes from those whose writing had been praised. One problem with these studies is that praise and criticism were contextually disembodied, simply given mechanically according to the group writers were assigned to, with no relationship to the quality of the writing, or teachers’ perceptions of students’ needs. Other work has recognised that to be effective praise needs to be credible and informative and that insincere praise is unlikely to encourage good writing (Brophy, 1981). This is particularly the case at early stages of the L2 writing process where it is suggested that premature praise may actually confuse students and discourage revisions (Cardelle & Corno, 1981). Studies of L2 students’ reactions to teacher feedback show that learners 188 F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 remember and value encouraging remarks but expect to receive constructive criticism rather than simple platitudes (Ferris, 1995). One feature which may influence patterns of praise and criticism in written feedback is teacher response style. Anson (1989), for instance, has argued that the ways teachers judge writing and define their role when giving feedback are influenced by their belief systems. Such beliefs are partly the result of personal constructs but also originate in the social context in which teachers work. He suggests that teachers typically respond to student writing in one of three ways. Dualistic responders focus largely on surface features and take the tone of a critical judge of standards. Relativistic responders attend almost exclusively to ideational aspects of the writing, often ignoring significant linguistic and rhetorical problems. Finally, reflective responders respond to both ideas and structure and attempt not to be dictatorial in their approach. Developments in teaching theory and research have perhaps moved many teachers away from a dualistic response style, seeing it as prescriptive and potentially damaging to the student’s writing development. Severino (1993) adds a socio-political dimension to this, pointing out that the nature of a teacher’s response can suggest a stance towards both linguistic and cultural assimilation in L1 and L2 contexts. Teacher response style may also be influenced by other factors, which can include the language ability of students, task type, and the stage at which feedback is given. Feedback offered at a draft stage will often be different from feedback on a final product, intended to perform a different function. Many teachers view feedback on drafts as more developmental and so offer more critical comments on specific aspects of the text, while feedback on a final product is likely to give a holistic assessment of the writing, praising and criticising more general features. Thus, any study of teacher written feedback must take into account the interplay between teachers, students, texts, and writing purposes and so consider written comments as ‘‘multidimensional social acts in their own right’’ (Sperling, 1994, p. 202). Participants and data Our written data consists of all the teacher written feedback given to six ESL writers from various language backgrounds on a 14-week full-time English proficiency course at a New Zealand university. Writing was an important aspect of this course, with about 2 h each week given over to teaching academic writing and another 2 h devoted to writing workshops. Feedback was collected for all students in two classes: one class preparing students for undergraduate admission (Class A) and the other preparing them for postgraduate studies (Class B). The teachers of the two classes were both experienced ESL writing instructors who had taught the course several times before. Three students from each class agreed to participate as case studies. In Class A, there were 15 students. The composition of the class was quite varied and included F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 189 students from Russia, Somalia, Vietnam, Korea, Indonesia, Japan, China, Taiwan, and Hong Kong. The case studies (all pseudonyms) were Maho, a 19-year-old female from Japan, Keith a 26-year-old male Taiwanese student, and Seng Hee, a 20-year-old female Korean student. Like the rest of the class, these students were low to mid intermediate level and came to New Zealand with the aim of studying for a first degree. Class B was more homogenous and consisted of 16 students, including 7 students from Thailand and 6 from China. They had all completed a first degree in their own language and were planning to enroll for postgraduate studies after this course. The case studies here comprised Liang from Taiwan and Samorn from Thailand, both female students in their 30s, and Zhang, a 27-yearold male from China. These students were all at a high intermediate to advanced proficiency level. All the feedback on all the case studies’ writing assignments was carefully documented and categorised. This amounted to 10 pieces of work for Class A and 7 pieces of writing for Class B. Three pieces of writing in both classes involved a feedback/revision cycle, consisting of the writing of a draft, followed by written feedback, and then a revised version in response to the feedback. Regular observations of the classes by one of the researchers, with particular attention given to the writing workshops, provided information on the context within which feedback was given. Interviews with teachers offered another perspective on this context, as well as another source of information on feedback practices. The teacher interview prompt sheet (see Appendix A) consisted of questions covering the teachers’ approaches to teaching writing and giving written feedback and their expectations of student behaviour after feedback. In addition, they were asked to describe an occasion where they felt that they had given very successful feedback and to pass one piece of advice about giving effective feedback to a new teacher of ESL writing. Teachers were also asked to conduct think-aloud protocols as they gave written feedback to the draft of one piece of writing for each case study participant. A retrospective interview was then carried out with the students within a day of revising the draft. Photocopies of the draft with feedback and the final version of the writing were used as a visual prompt during the interview (see Appendix B). To ensure content reliability and combat researcher bias, triangulation and respondent validation were included in the research design. Triangulation involved obtaining as many different perspectives on the data as possible. These different perspectives came first from the different sources of data: the teachers, the students, and the researcher and from a triangulation of methods, as data were collected through interviews, questionnaires, analysis of texts, observation of classes, and verbal reports. Respondent validation, or ‘‘member checking’’ (Lincoln & Guba, 1985), involved allowing participants in the research access to data and seeking their input and their evaluation of its authenticity to correct researcher bias. All observation notes were commented on by the two class teachers to help validate the interpretations and minimise misrepresentations. 190 F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 Analysis and categorisation scheme While there have been a number of models suggested for classifying teacher comments, many of these have often focused on contrasting large-scale areas such as ‘‘content’’ vs. ‘‘form’’ (Searle & Dillon, 1980), ‘‘local vs. global’’ issues (Zamel, 1985), and ‘‘high order vs. low order’’ concerns (Keh, 1990) and have not addressed teachers’ aims. Important exceptions are Ferris’ models (Ferris, 1997; Ferris, Pezone, Tade, & Tinti, 1997). However, while these are perhaps descriptively more useful, they contain rather overcomplex lists of text variables, which may be too detailed to be used by teachers wanting to examine their own feedback. More problematically, the schema tends to confuse pragmatic and formal criteria and contains considerable overlaps. Thus, coders are asked to categorically distinguish questions which give information and those which ask for it (Ferris, 1997), for example, and to identify comments where teachers are asking for ‘‘known information’’ and where they are rhetorically seeking to ‘‘spur the student to further thought’’ (Ferris et al., 1997, p. 164). We feel that the constraints acting on teachers when providing feedback make it unlikely that they deliberate between, say, ‘‘asking for unknown information’’ or ‘‘making a request’’ in their end comments. Nor is it entirely clear how far either the analyst or the reader can determine which meaning is intended. We argue instead that teachers constantly seek to engage with both the student and the assignment through a limited number of overarching pedagogically and interactionally effective comment functions. We developed our coding categories inductively through our reading of the essays and drafts. First, we identified individual ‘‘feedback points’’ as single written interventions that focused on a particular aspect of the text. The two researchers, working independently, had little difficulty in determining where one point ended and another began. Even long comments could usually either be clearly identified as carrying a single pragmatic force or as breaking down into a series of related speech acts. The few occasions where we could not agree were resolved by discussion. While we acknowledge that all feedback, including both ‘‘in-text’’ and ‘‘end’’ comments, have the potential to affect students powerfully, for this study we decided early on to ignore in-text points which occurred in the margin or the body of the text. In fact, many of these were simply symbols or codes which focused on language inaccuracies and corrections. We chose to examine only end comments, those written either at the end of the essay or on a separate feedback sheet, or on both, since these were longer, more substantive, and more discursive remarks on the student’s writing overall. In total, 495 feedback points were examined, totaling 4700 words from 51 student essays. Having identified the feedback points, we then worked independently to code each one according to its main evaluative or pedagogic purpose of praise, criticism, or suggestion. All comments were, thus, double-coded for reliability, and while we sometimes disagreed on whether a feedback point was principally a suggestion or a criticism, the few discrepancies were quickly resolved. The F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 191 coding system we evolved seemed potentially useful as it effectively characterised the need for teachers to both retrospectively respond to the completed draft and to proactively address the students’ future work. The meanings expressed in the end comments not only centre on the key functions of feedback, but also provide a more straightforward and readily identifiable pragmatic and content focus than Ferris’ categories. They indicate, for instance, that ‘‘making a request,’’ ‘‘giving information,’’ ‘‘making a grammar comment,’’ etc., are essentially means of praising, criticising, and suggesting. This is not to say that there are no overlaps between our three categories, particularly between criticisms and suggestions. Embedded in every suggestion is an assumption that the original text requires improvement, and this criticism can be more or less explicit in the way the feedback is expressed. Thus, in (1) we have a fairly clear suggestion for revision, while in (2) the same teacher has chosen to express her comment more forcefully as a criticism: 1. Try to express your ideas as simply as possible and give extra information. 2. There is no statement of intention in the essay — what is the purpose of your essay and how are you going to deal with it? You are not giving me any direction. The recipient of the feedback in (2) clearly has to do more interpretive work to recover the implications of the comment, and presumably the teacher hopes that the careful student reader will be able to unpack the intended advice it carries. However, this may not always occur, particularly if the comment is couched in general terms: 3. Repetition of sentence beginnings does not always work when too many start the same way. In other words, we see criticism and suggestion as opposite ends of a cline of expression ranging from a focus on what was done poorly to measures for its improvement. We distinguished suggestions then, partly through the fact that they contained a retrievable plan for action, a do-able revision of some kind, which either addressed the current text (4) or extended this to future writing behaviour (5): 4. You need to decide on your main point at some stage in the process and connect everything to that. 5. For your next essay, I suggest you use written references (books, etc.). While suggestions were more or less directive in this way, a second way of identifying them was through their surface structure. Over 3/4 of all suggestions were expressed by a limited number of explicit formulae, principally including the modals of assumption and obligation need to, could, and should, hypothetical would, and the verb try. Comments which lacked any of these features were generally criticisms. 192 F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 The important point to make here is that teachers face choices when responding to student written work, selecting from available options which carry very different pragmatic force. We will argue in what follows that these decisions are often based on a desire to negotiate interactions which recognise both the learner’s struggle to make meaning in a foreign language and the fragile intimacy of the teacher– student relationship. Teacher acts in end comments Our results show that 44% of the almost 500 feedback points were related to praise and 31% were critical. Only a quarter were comments offering explicit suggestions (Table 1). These findings are immediately interesting as they contradict the work which claims that positive comments are rare in feedback (e.g., Connors & Lunsford, 1993). Our experience is that teachers frequently use praise and regard it as important in developing writers (see also Bates, Lane, & Lange, 1993). In interviews, both teachers showed that they were very aware of the potential effects of both positive and negative feedback. Joan (Teacher A) said that her general principles for giving written feedback were to find something positive to say and to look for the most important and most generalisable problems to comment on. Nadia (Teacher B) also reported that she tried to guard against the overuse of negative comments. She felt that it was necessary to show both the positive and negative parts of the writing, but stressed that the positive was ‘‘incredibly important’’ since ESL writers were very insecure. She tried to be more positive in her feedback on a final piece of writing, whereas on a draft she felt that she could be harder because the students were aware that to get a good grade later, they would need to ‘‘fix up’’ aspects of their writing which were weak. One aspect of Table 1 which is immediately striking is the difference in the total number of comments given by the two teachers, and especially in the greater number of suggestions made by Joan. Although a similar number of pieces of writing were commented on by both teachers, Joan made 109 suggestions, whereas Nadia made only 15. This can be partly explained by the different stances of the two teachers. In her interview, Joan stressed her dual role when giving feedback ‘‘as facilitator and provider of knowledge,’’ pointing out ‘‘which aspects are good and which can be improved.’’ In contrast, Nadia reported that she preferred to respond globally rather than deal with discrete points. In addition, the way they gave feedback possibly affected the number of suggestions. Joan used a feedback sheet Table 1 Teachers’ use of feedback acts Joan Nadia Totals Praise Criticism Suggestion Overall 160 (42%) 58 (52%) 218 (44%) 114 (30%) 39 (35%) 153 (31%) 109 (28%) 15 (13%) 124 (25%) 383 112 495 193 F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 and tended to give feedback on each of the points it specified (organisation, introduction, conclusion, language accuracy, etc.). Nadia, on the other hand, wrote her feedback directly on the student paper as end comments which may have led to her making more holistic and less specific observations. We have already noted that feedback on developmental writing may have a different purpose to that given on final essays, and so we also decided to look at comments according to whether they were given in a draft or final version. Three assignments involved students writing two drafts, the second in response to teacher written feedback. The distribution of praise, criticism, and suggestions across drafts is given in Table 2. It is interesting to note that although criticism and suggestion were fairly evenly distributed between drafts and final versions, nearly three quarters of all praise was reserved for final drafts. Interviews with the two teachers revealed that Nadia felt happier offering critical comments on drafts where there was potential to improve them, and Joan was uncomfortable making critical comments on drafts without appending a positive comment. In final versions, it seems that praise was extensively used to motivate the students in their next writing. There also appears to be similar thinking behind the high use of suggestions as many of these moved beyond specific problems in the current text to provide more general advice on writing. When we consider the data in Table 2 vertically, looking first at the distribution of teacher acts in the draft and then in the final versions, we see more clearly that although the type of feedback the students received most often overall was praise, the feedback they received most on first drafts was criticism. Our analysis of these two teachers’ responding practices also sought to address the principal areas of feedback focus by categorising the general aspects of the writing they chose to target in their comments. We found that comments addressed five main areas: the students’ ideas, their control of form and mechanics, their ability to employ appropriate academic writing and research conventions, the processes of writing, and global issues, relating to the entire essay. Table 3 shows that comments overwhelmingly addressed the ideational content of the writing. The teachers focused their praise mainly on ideas (64%) and were much less likely to praise either formal or academic aspects of the texts. Criticisms also tended to address this aspect and comprised over 43% of all the criticisms given, with less than a quarter concerned with formal issues. Suggestions focused evenly on ideas and academic concerns and slightly less on formal language-related issues. Although it may appear that the teachers were giving Table 2 Distribution of teacher feedback acts in drafts and final copies Praise Criticism Suggestion Totals Draft Final Total 33 40 34 107 93 34 39 166 126 74 73 273 (26.2%) (54.0%) (46.6%) (39.2%) (73.8%) (46.0%) (53.4%) (60.1%) (100%) (100%) (100%) (100%) 194 F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 Table 3 Focus of teacher feedback acts (%) Praise Criticism Suggestion Overall Ideas Form Academic Process General Totals 63.8 43.8 36.3 50.7 13.8 23.5 23.4 19.2 13.3 30.1 36.3 24.3 2.3 2.6 4.0 2.8 6.8 0.0 0.0 3.0 100 100 100 100 priority to meaning issues, it is worth reiterating that we focused only on the summary comments in this study and that a great deal of in-text feedback actually had a language focus (see Hyland, F., 1998). Mitigation in teacher end comments While it is an important pedagogic resource, teacher written feedback is also an evaluative genre in which student writing is judged and pronounced on. Evaluation always carries with it the seeds of potential friction because criticism can represent a direct challenge to a writer and undermine his or her developing confidence. Praise, too, also carries risks, for while it conveys support and interest, it can also damage an open relationship between the teacher and student as it implies a clear imbalance of authority. As every teacher knows, responding to student writing entails more than deciding whether to comment on form or content; it involves delicate social interactions that can enhance or undermine the effectiveness of the comment and the value of the teaching itself. This is one reason why teachers may choose to compliment the ideas in a student essay or opt to present a criticism in the form of a suggestion. However, the response choices available to teachers are not limited to these broad macropurposes. It is clear from our data that when expressing judgements, the two teachers were aware of the affective, addressee-oriented meanings their comments conveyed. As a result, baldly negative comments such as ‘‘Poor spelling’’ or ‘‘Referencing is inadequate’’ were rare. In fact, 76% of all criticism and 64% of suggestions were mitigated in some way. Praise was presented baldly 75% of the time but was itself widely used to tone down the negative effect of comments. Redressive strategies principally involved paired-patterns, hedges, personalisation, and interrogative syntax, and often an act included at least one of these strategies. The ways that these were distributed across different categories are shown in Table 4. Paired act patterns One of the most obvious features of our data was the frequency with which these teachers combined their critical remarks with either praise (20%), sugges- F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 195 Table 4 Acts mitigated by different strategies Praise Criticism Suggestion Overall Paired acts Hedged Personal Interrogative Unmitigated 68 67 35 156 26 79 44 135 5 17 14 42 2 19 5 23 117 (53.7) 24 (15.7) 36 (29.0) 177 (35.8) (31.2) (43.8) (28.2) (31.5) (12.0) (51.6) (35.5) (27.3) (2.3) (11.1) (11.3) (4.9) (1.0) (12.4) (4.0) (4.7) Acts can be mitigated by more than one strategy simultaneously. tions (15%), or both (9%). One-fifth of the criticisms were accompanied by praise, the adjacency of the two acts serving to create a more balanced comment, slightly softening the negativity of the overall evaluation: 6. Vocabulary is good but grammar is not accurate and often makes your ideas difficult to understand. The order is OK, but the problem with this essay is the difficulty of finding the main idea. You used the information in the diagram well. Although you did not mention the expansion of the earth’s crust. Here then teachers seem to syntactically subordinate criticism to praise by preceding a negative comment with a positive one, a strategy we found later to be all too obvious to the students themselves. Fifteen percent of criticisms were linked to suggestions, expanding what might be seen as a blunt criticism into a proposal for improvement, thus adding a more effective pedagogic and interactional dimension: 7. This conclusion is all a bit vague. I think it would be better to clearly state your conclusions with the brief reasons for them. This is a very sudden start. You need a more general statement to introduce the topic. This essay tends to waffle on a bit, I’m afraid. Try to make it much tighter and clarify your thoughts. Once again, the full force of the criticism is assuaged by the second part of a pair. As the above examples indicate, suggestions could focus on the students’ texts and proposals for revisions, but some given on both drafts and final versions of writing covered general principles and extended the suggestions to future writing behaviour, as in (8): 8. Maho, as I said on your first draft, a lot of this essay is about your learning history and, therefore, not directly relevant to the topic. At least you haven’t shown how it is relevant. At university, you must answer the question you choose and keep on the topic. 196 F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 This example is also interesting for its dialogic nature or intertextuality (Bakhtin, 1981) expressed through the back references to previous feedback. There were many examples in this data of feedback referring back to previous advice, to advice given orally during writing workshops, and also to points made by the teacher during classroom teaching sessions. Such examples illustrate the importance of context for a complete understanding of the analysis of feedback. Following praise – criticism and criticism – suggestion, the most common pattern in our data was the praise– criticism –suggestion triad. This strategy serves to both mitigate the potential threat of the criticism and to move the students towards improving either their current text or their writing processes more generally in the longer term: 9. References very good. Two small problems. (1) Bibliography (at end of essay) — include initials of author. (2) Be careful about referencing inside the essay. Avoid said. Interesting content, but difficult to understand. I think you need to ask for help from flatmates, classmates, friends, me to read your writing and see if they can understand it. For the most part, praise was less specific than criticism and frequently more cursory than developmental. In fact, previous work on compliments has suggested that these tend to be highly formulaic, in terms of both their grammatical structure and limited range of vocabulary (Johnson, 1992). However, while our data includes some comments which focused quite closely on what was good, suggestions tended to be pedagogically more useful and often functioned to narrow the issue considerably in praise –suggestion pairs: 10. Good movement from general to specific. But you need to make a clearer promise to the reader. This is a good essay but you have to expand your ideas. In sum, criticisms of student writing in our data were frequently accompanied by either praise or suggestions which acted to fulfil both pedagogic and interpersonal functions for teachers. Hedges A second mitigation strategy was the use of hedges. Not surprisingly in such an evaluative genre, hedges were widely used to tone down criticisms and reflect a positive, sympathetic relationship with student-writers. While hedges have both an epistemic and affective function (Hyland, K., 1998), their principal purpose here was not to suggest probability, but to mitigate the interpersonal damage of F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 197 critical comments. Our findings here differ considerably from Ferris’ (1997) study of the feedback of one experienced ESL teacher, in which only 15% of comments contained hedges. This may be because we recognised more features as hedges, or it might relate to cultural or individual differences in feedback practices. In our data, hedges occurred frequently in end comments and comprised mainly modal lexical items, imprecise quantifiers, and usuality devices such as often and sometimes. The following comment, for example, is structurally a criticism – suggestion pair, but the teacher has further drawn the sting from the comment by including a number of lexical softeners: 11. Some of the material seemed a little long-winded and I wonder if it could have been compressed a little. Here the comment is made more tentative through the imprecise weakening devices some and a little, the lexical verbs seemed and wonder, the modal could, and the use of the hypothetical form. The response is weakened further by being couched in personal terms, a strategy we will discuss further below. Hedges largely occurred in evaluations of content, where one might expect the interpersonal effects of direct criticism to be potentially most harmful, but they were also found in other categories of comment. They also appeared to be most used to mitigate the full force of stand-alone acts of criticism: 12. This is actually a little bit too long. Your conclusion was a bit weak. The essay is rather middle-heavy. There is possibly too much information here. Hedges not only occurred in criticisms. Both teachers also used them to weaken the impositions which a direct suggestion may make on the reader: 13. Tomoko wrote well on this, you might ask to read hers. You could still make the promise to the reader clearer. You might also need to study the use of ‘ing’ and ‘ed’ forms in complex sentences. It might also be good to change the order of your paragraphs/ideas. Perhaps surprisingly, hedges also often occurred in statements of praise. While common, praise was rarely fulsome and was at times very faint indeed: 14. Although parts of this essay have improved, they seem to be mainly surface level adjustments. Fairly clear and accurate. 198 F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 You show a reasonable understanding of this. Mostly fairly good. It seems here that praise is being used less to compliment the writer on an aspect of his or her essay than to call attention to some weakness. Hedged praise, in other words, might work for the writer as a rephrased criticism, simply prefacing a criticism or signaling a problem in a way which is less threatening to the teacher –student relationship. Personal attribution A third strategy to soften the force of criticisms was to signal it as reflecting a personal opinion. As we noted above, judging another’s work is always an inherently unequal interaction because the power to evaluate is nonreciprocal and lies exclusively with the teacher. By expressing their commentary as a personal response, however, teachers can make a subtle adjustment to the interactional context and perhaps foreground a different persona. It allows them to relinquish some of their authority and adopt a less threatening voice. In other words, personal attribution allows teachers to react as ordinary readers, rather than as experts, and to slightly reposition themselves and their relationship to the studentwriter. This can be seen in the following examples: 15. I’m sorry, but when reading the essay, I couldn’t see any evidence of this really. Perhaps you should have given me your outline to look at with the essay. I find it hard to know what the main point of each paragraph is. My concern in this essay is that you introduce several terms in the introduction but do not provide a definition for any. Specifying oneself as the source of an opinion can qualify its force by acknowledging that others may hold an alternative, and equally valid, view (Hyland, 2000a; Myers, 1989). Thus, in (15) there is a suggestion that another reader may see the required evidence or find the main point of each paragraph. The personal expression of criticism, then, reminds the reader that the comment carries only the view of one individual, thereby conveying the limitation of the criticism. The decision of writers to soften censure by rhetorically announcing their presence in the discourse was also employed to phrase suggestions. Representing a comment as their individual opinion rather than an uncontested necessity helped to mitigate what might otherwise be seen as a command: 16. I still believe a major effort to read English would improve your grammar. At university, I suggest you try to use a computer with a spelling checker because lecturers will not tolerate bad spelling. F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 199 Interrogative syntax The final form of mitigation employed by these two teachers in their summary comments was to construct the criticism in interrogative form. Questions are a means of highlighting knowledge limitations and can be used to weaken the force of a statement by making it relative to a writer’s state of knowledge. While they generally seek to engage and elicit a response from the reader, questions also express the writer’s ignorance or doubt and, therefore, can mitigate the imposition of a suggestion or a criticism: 17. The first two paragraphs — do they need joining? Did you check your spelling carefully? Why not make a spelling checklist of words you often get wrong and use this before handing in your final? I think you have selected appropriate ideas to include. Why do you want to add more ideas? In addition, questions are also useful when one wishes to protect oneself or one’s reader from the full effects of what might be considered serious allegations. Interrogatives undermine a categorical interpretation of the underlying proposition, and because of this they can, in some cases, be seen as the teacher’s attempt to withhold full commitment from the possible implications of a statement, an important consideration when the issue is one of plagiarism or the extent to which another person was involved in writing or editing the essay: 18. You only mention Ward once in the essay. Are all the other ideas your own? You need to make it clear which are yours and which are hers. Have you used quotations here? Some of it sounds like it might be. Did you get some help with the editing? Mitigation: motivations and miscommunications Motivations Why did the teachers use so many mitigation devices when giving feedback? We explored their motivations partly by examining the talk-aloud protocols they conducted when responding to the case study subjects’ assignments and partly by examining the interviews carried out with both teachers at the beginning of the course to discover their beliefs and reported practices about feedback. The two teachers took slightly different roles when responding to student texts. Joan reported that she sought to stress the ‘‘most important or most generalisable problem’’ in her feedback and that she tried not to overwhelm the students by criticising all their problems. She mentioned an experience on a 200 F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 previous course which had affected the way she gave feedback, making her less willing to be directly critical: I had a Korean student who was kind of a fossilisation problem, I guess. And her writing was just full of errors and like you didn’t even have paragraphs and it was very short. On the very first test, I think I made some criticisms. . .and she wrote in her journal that she found this very devastating and ‘‘please try and encourage me’’ and so after that I modified my feedback to try and be more positive. I mean, I had been positive but I felt it was my duty to point out that there were major problems here. I mean, it’s hard sometimes to get a balance between being a realist and being positive. But once she told me that, I made a conscious effort. Nadia first suggested that she viewed her role when responding to writing as ‘‘a reader rather than a know-it-all teacher.’’ However, while this was her ideal, she was not always able to escape from a more authoritative role and was sometimes directive in her approach. Having elaborated her stance, she said to the interviewer, ‘‘Am I saying the right things or not? I’m probably going contrary to all kinds of research.’’ She, therefore, seemed to be aware that there were more and less appropriate ways of giving feedback and that a too dominant teacher role might be considered unacceptable. Nadia had previously worked on a first language writing course which had emphasised the importance of not appropriating student texts, and this agenda was obviously in her mind when both responding to interview questions and giving feedback. The topic of ownership of student texts has been an important and prominent topic in L1 writing research (see, for example, Brannon & Knoblauch, 1982; Onore, 1989). The arguments concerning appropriation are summarised by Knoblauch and Brannon (1984, p. 118), who have argued that writing can be ‘‘stolen’’ from a writer by the teacher’s comments. They suggest that if students follow directive feedback too closely, they may develop neither their cognitive skills nor their writing abilities, but merely rewrite texts to reflect their teachers’ concerns. Teachers’ fears of text appropriation have also surfaced in the ESL writing classroom. However, there has been some debate about their relevance and validity, especially for L2 writers. Reid (1994) has claimed that if they were overconcerned about appropriation, teachers might fail to give L2 students the direct and concrete help they needed. Nadia’s stance suggests that she was aware of the issue of ownership of texts, and the mitigation of her feedback may have been a way of toning down what might be seen as over-directive interventions in the students’ writing. By couching comments and criticisms as tentative suggestions, both teachers may have been attempting to intervene with clearer consciences, seeking to avoid demotivating students with negativity and removing the potential disquiet created by seeming to take over their texts. Another reason for mitigating the force of comments was to minimise the possible threat which criticism carries for the ‘‘face’’ or public self-image of F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 201 students. This was particularly evident in cases where teachers dealt with plagiarism, a very sensitive issue for feedback and something teachers are often unwilling to address directly. Teachers may also be aware of the literature that suggests that plagiarism is, at least partly, a western cultural concept (e.g., Pennycook, 1996; Scollon, 1994). Scollon (1994), for instance, discusses the ambiguity of the concept of plagiarism for nonnative writers unfamiliar with the individualistic authorial role expected in English academic writing which demands an inseparable link between ideas and those who first presented them. However, while students from cultures with more collectivist discourse practices may regard the nonattribution of authorship more positively than university handbooks, ESL teachers often feel that it is important to address these concerns when giving feedback. Not wishing to seem culturally insensitive, however, they often appear reluctant to do this directly. Both teachers in this study disliked approaching students head-on over issues of plagiarism. In two think-aloud protocols conducted when giving feedback on assignments, Joan debated how to address this issue and make the students aware of plagiarism. She finally decided on indirectness: It doesn’t sound like her words — I hate accusing people of plagiarism, but when you think it is, what do you do? Joan wrote as her feedback — Where did you get this information? Have you used quotations? — and considered this interrogative form ‘‘a subtle way of saying it.’’ Unfortunately, the student’s retrospective interview revealed that she failed to identify these as rhetorical questions and to detect the underlying criticism and the implicit suggestion. The offending text remained in her final draft. Joan again felt uncomfortable with this issue in another case and debated how to tell Maho that she had plagiarised: Uh so, bother — do I accuse her of plagiarism? — I can ask her where she got this information from anyway. Once again, she used the formula, ‘‘Where did you get this information from?’’ in her feedback rather than directly confronting the issue. Nadia, when confronted with a case of plagiarism of an entire essay, was also unwilling to be direct. She responded with a very indirect comment comprising a personalised and hedged criticism followed by a second considerably hedged criticism: I am afraid that this may not be your own work. You may have gotten some/ considerable help with it. Perhaps unsurprisingly, this failed to open a dialogue with the student on the topic. A final reason for the use of mitigation strategies is that teachers often see them as a means of helping to maintain or develop good relationships with their students. As we have seen, one way of taking the sting out of critical remarks was through the use of positive comments, but students did not always welcome empty remarks, and their reception of praise varied considerably. Some thought 202 F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 such feedback served no useful function. A student in Nadia’s class, Zhang, for example, believed positive remarks were ‘‘useless’’ unless they were backed up by ‘‘serious’’ comments that he could act on: Because the teacher gives only a few words — It’s OK or it’s interesting — I think it’s useless. [. . .] It is difficult for the teacher to be serious all the time, but in fact the teacher should try their best to be more serious. The students, they think very much about the teacher’s feedback. If the teacher always not serious, the students will be feel very disappointed. This view of positive comments was reinforced by another student in Nadia’s class, Mei Ling, who said she discounted positive comments because ‘‘I want to know my weaknesses most.’’ In her interview, she spoke against positive feedback as a waste of time; what she wanted was what she termed ‘‘negative’’ feedback highlighting her problems. For her, positive comments were insincere and, therefore, worthless: Sometimes maybe the teacher doesn’t mean it, but they just try to encourage you. [. . .] Because there is always ‘‘but’’ after the positive. Sometimes the teacher just tries to find something good in my essay and then may be that strength is not the main point. Her observation, incidentally, is an astute piece of informal discourse analysis, as our data shows that most of the praise –criticism pairs were linked by a contrastive connector. However, the teachers were also right in their assumption that positive comments could be motivating for students. Maho emphasised the importance she attached to them: If teacher give me positive comments it means I succeed. Nadia’s student Samorn also described how a lack of positive comments affected both her attitudes to writing and her reception of feedback: . . .if feedback is not so good, I mean that teacher criticise many mistake I have, then I feel — ‘‘Oh I don’t like writing.’’ I am very interested in teacher’s comments every time. I like to read it and when I read it and if it says ‘‘it’s good but your problem is grammatical problem,’’ then I will turn back to see how many mistakes I have. But if the comment is very bad and maybe not good enough, maybe I’ll stop for a while and keep it and take it out and look at again later. These student perspectives on positive comments suggest that they need to be used with care by teachers, rather than just included to make critical comments more palatable. Miscommunications While we did not initially set out to look for cases of miscommunication, it soon became clear during the student interviews that they were often unable to F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 203 understand the teachers’ mitigated comments. In fact, teachers did sometimes reflect aloud whether students would understand their more indirect statements. For example, Joan debated this possibility as she responded to an essay although she ultimately let the comment stand: The main idea is clear — but the organisation of sentences in the introduction — and conclusion — is confusing? No. Is chaotic? Ha — but the organisation of sentences in the introduction and conclusion — is jumbled? Um could be improved? err — could be improved — sometimes students don’t understand that as a criticism but he should. While we have no quantifiable data to determine the full extent of this problem, each case study provided several examples where students failed to understand, or only partly understood, such mitigated comments. This study allows us to do little more than point to this as a potentially serious issue, but we believe that it opens interesting possibilities for more detailed research. In this final section, we draw on the transcripts of the retrospective interviews and the teachers’ protocols. This preliminary data strongly suggests that teachers may sometimes forget that students are reading their feedback in a foreign language and that being more indirect and ‘‘subtle’’ may actually result in significant misunderstandings. Our first example comes from Nadia’s response to Zhang’s assignment where she wanted him to define a term which he used throughout the essay. She drew his attention to this with a bald criticism followed by an interrogative suggestion: My concern in this essay is that you introduce several terms in the introduction but do not provide a definition for any. I should like to know more about ‘macroscopic’ — is this what you are talking about in your conclusion? However, in his final draft of the essay, the term was still being used without a definition, and Nadia commented that this aspect ‘‘still hasn’t changed’’ in this version. Zhang was surprised and somewhat put out to read this comment and said that he ‘‘didn’t understand what she wanted’’: R Z R Z R Z R Z R Z So why do you think your teacher made that comment? I think we have err, maybe I understand this word, no. I can’t use this word correctly. The meaning I do not understand very well. Is that what she’s telling you? That she thinks you don’t understand it? No. What’s she saying? She did not say anything about that. She mentioned macroscopic — know more about it. Yes well it seems to me that this word means overall, but I don’t know whether it’s right or not. Now here she’s written still hasn’t changed (final copy). What do you think she’s saying there? I don’t know. 204 R Z F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 Do you think she wanted you to put in some definitions here of words like macroscopic? I don’t know, in fact I didn’t understand what she meant. Clearly, Zhang did not simply refuse to respond to this comment on purpose: ‘‘I didn’t mean to ignore it, but I did not understand it. I think the teacher is always right.’’ Other examples of miscommunication were found in all the case studies. For example, Joan noted a problematic statement in the final paragraph of Keith’s assignment and discussed this in her protocol: (Joan reading) ‘‘based on above data, we can perceive the higher sales revenue creates more profits’’ — of course. Now should I say something about this — I’ll ask isn’t this statement obvious/always true? — I hope he doesn’t mind the suggestion. The retrospective interview revealed that Keith failed to understand the negative implication of the question: R K R K R K R K Right — Why has she written isn’t this statement obvious/always true? I don’t know. Did you do anything with that comment? No. Do you know what she means by it? Isn’t this obvious/always true? Just a question I think. She’s just wondering? Yes wondering. But. . .I thought this is true. In another part of the assignment, Joan felt that Keith was repeating himself and tried indirectly to alert him to the fact that certain information was superfluous with a hedged criticism: It seems funny to have this repeated — oh so this second part is like a summary — this just seems like a summary — it’s not necessary. I might just write — The second section seemed like a summary of the previous information. However, this comment was misinterpreted as a positive comment and served to reinforce this strategy: R K R K Why did you change this part? Because here — Your second section seemed like a summary. Because in my proposal, this includes two parts, so I think this is very necessary. At the end of part one I make, like Joan says, a summary. So also I have put in another summary in here. Right so you now have short summaries at the end of each part? Yes. F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 205 A further example comes from Class B. During her protocol, Nadia gave the following comments on Samorn’s conclusion to her essay comparing a free market and planned economy: ‘‘In a free market economy, there are more productive efficiency than in a planned economy and consumers are happier for they can choose and get the goods they want.’’ Ha she clearly knows which one she wants, good. . . .Still the conclusion is a bit abrupt. I might just write — the conclusion may be a bit abrupt — you could re-state some of the main points. Samorn, however, misinterpreted this hedged criticism – hedged suggestion pair as a negative comment on her stance in the conclusion, as this extract from her interview shows: R S Right, where it says here, ‘‘The conclusion may be a bit abrupt, you could restate some of the main points in the conclusion’’ — do you do that at all? Yes, I think I do. In the conclusion, I change it from my first draft. My first draft I say that the free market economy is better than a planned economy, but this one I will not say like that, I will say I can not tell which is one is better. And I know that in fact I shouldn’t have said that which one is better because I cannot say that. We have already discussed Joan’s reluctance to directly accuse Maho of plagiarism by couching her comment as an interrogative criticism: ‘‘Where did you get this information from?’’ This phrasing meant the meaning was impenetrable for Maho, however, who took the question literally as a request for a reference, as can be seen from her interview transcript: R M R M R M ‘‘Where did you get this information from?’’ Did you change anything to show where you got the information from? Not yet and I wonder where can I put all those information, references? Right. So you weren’t sure how to do the referencing? Not quite sure. So you’ve left it out? Yes. Miscommunication can occur even when teachers seek to address students’ requests for particular kinds of help. Maho, for example, wrote on her essay cover sheet that she wanted feedback on her ideas and also some suggestions for new ideas. This was how Joan responded in her protocol: There are — no — you is more personal perhaps — You have interesting ideas here and seem to have thought deeply about the ideas — mmm — I think I might comment on — she wanted more ideas? — I might comment on that she 206 F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 has selected ideas well — I think you have selected good — maybe some of those actual examples weren’t very good appropriate ideas to include — and I might just ask her why do you want to add more ideas? Thus, Joan first considered a comment which gave hedged praise, moved to personalised and hedged praise, and finally ended up with a criticism mitigated by an interrogative. In an interview, Maho revealed that she was confused by the indirectness of this comment and reinterpreted it as a criticism of what she saw as her major problem, that of organisation: R M R M R M ‘‘You have some very interesting ideas here and seem to have thought deeply about the ideas. I think you have selected appropriate ideas to include. Why do you want to add any more ideas?’’ Why do you think your teacher made this comment? Because my ideas not concentrate on one point. Just this idea’s about that one and this idea’s about this one here. Something like that. I mean she means I think. Right so you think that she means you have different ideas in different places. Yes not well organised. From this comment you think that? Yes. We have attempted to show from these extracts that while teachers often have laudable interpersonal and pedagogic reasons for mitigating their feedback, tentative comments have the very real potential to cloud issues and create confusion. While further research is needed to explore this in greater detail, it seems that mitigated criticism was most opaque to students and a source of particular confusion, especially where it was phrased interrogatively and not coupled with an explicit suggestion for revision. Indirectness frequently seems to be counter-productive to the aim of clearly conveying the point the teacher wishes to make and is often reinterpreted by students according to their own writing concerns and agendas. In other cases, failure to understand implied criticisms or toned down praise leads the student to revise aspects of the text which are not problematic. Our data clearly illustrate that indirectness can result in communication problems, and this is an aspect of feedback which needs to be investigated more directly by future studies. Some conclusions and teaching implications This paper focused on an important aspect of teacher feedback: the summary comments at the end of student assignments, and considered them F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 207 in terms of their functions as praise, criticism, and suggestions. Critical analysts have consistently argued that patterns of language use contribute to the relationships which help structure social positions, and the two teachers discussed here seemed very aware that their responses to student writing had the potential to construct the kinds of relationships which could either facilitate or undermine a student’s writing development. They recognised that offering praise and criticism expresses and confirms the teacher’s right to evaluate a student’s work and, as a result, sought to blur the impact of this dominance. By combining these acts into patterns of praise –criticism, criticism – suggestion, and praise – criticism – suggestion, and through use of hedges, question forms, and personal attribution, they sought to enhance their relationship, minimise the threat of judgement, and mitigate the full force of their criticisms and suggestions. It is clear, however, that while responding to student writing is an important element of the teacher’s role, it is also a practice that carries potential dangers and requires careful consideration. We have seen that despite the best intentions of these teachers and their desire to respond positively and effectively, the effect of their mitigation was often to make the meaning unclear to the students, sometimes creating confusion and misunderstandings. This result, incidentally, contrasts with Ferris’ (1997) finding that hedged comments were more likely to lead to positive revisions than those without hedges. Presumably, the hedges did not directly encourage greater revisions, but were included in comments that were more directly usable. That is, teachers are more likely to hedge suggestions and criticisms (both of which implicitly or explicitly invite revision) than positive comments (many of which cannot be acted on). Future studies might investigate such links between ‘‘usable’’ hedged and nonhedged feedback and revision practices, using both analyses of student essays and interview data, so that we could learn more about the effects of hedges. While our analysis represents the feedback comments of only two teachers and should, therefore, be treated cautiously, we believe these mitigation patterns are used more widely and, thus, our findings have potential application for L2 pedagogy. The kinds of miscommunication we have identified suggest that it may be a good idea for teachers to look critically at their own responses and to consider ways of making them clear to students. One aspect they could consider more closely is the number of mitigation devices they use. Although teachers need to guard against being over-directive, sometimes it may be necessary to deal with problems and possible solutions quite frankly. This is especially important with learners of low English proficiency since they may be less familiar with indirectness and fail to understand implied messages. Nor should we overestimate the ability of more advanced learners to recover the point of our remarks, as hedges are often invisible to L2 readers (Hyland, 2000b). Indirectness, in other words, can open the door to misinterpretation. More research that employs retrospective interview techniques to look at how students actually understand indirect feedback is needed. 208 F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 Our data also strongly suggests that misinterpretations were common when dealing with plagiarism, due to the teachers’ unwillingness to address this issue directly. However, while hedging our comments on such sensitive topics may cause us less discomfort and help preserve the student’s face, we also have to think about his or her future participation in academic environments that invariably have exacting standards of attribution and referencing. Admonishments about plagiarism are common in style guides and academic handbooks, and most universities tend to treat the issue as a punishable offence. However, while we may be critical of these official practices or embarrassed by raising the issues in our comments on the work of particular students, shrouding the issue by forcing the reader through a maze of indirectness serves little purpose and fails to assist L2 learners seeking admission to further study. Like other cultural and rhetorical conventions of academic communities, we should address these issues clearly and directly, both in our classes and our feedback. At the same time, our discussion has shown that students vary considerably in what they want from their teachers in the form of feedback. Some students value positive comments very highly while others simply discount them as merely mitigation devices, so we need to take care when making positive comments. In particular, these may need to be specific rather than formulaic and closely linked to actual text features rather than general praise. Most importantly, praise should be sincere. Students, as we have seen, are adept at recognising formulaic positive comments which serve no function beyond the spoonful of sugar to help the bitter pill of criticism go down. In sum, when we pick up a pile of student essays to mark, we do not approach them with a tabula rasa. We have in our minds a stock of tried and tested phrases to choose from which relate in various ways to our own experiences of feedback, the kinds of teachers we are, and what we are trying to achieve. We alter these to fit specific students and their needs and personalities. This means that our comments go far beyond simple decisions to address form or content or to praise mechanics or criticise organisation. We are generally acutely aware of the importance of feedback in both providing helpful advice on our students’ writing and in negotiating an interpersonal relationship which will facilitate its development. The ways that we frame our comments can transform students’ attitudes to writing and lead them to improvements, but our words can also confuse and dishearten them. We hope that the description we have offered here may encourage teachers to reexamine their feedback to ensure it is clear and constructively helpful to students. Acknowledgments We wish to thank two anonymous reviewers for their careful reading of our paper, their clear (and often unmitigated) criticisms, and their very helpful suggestions. F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 209 References Anson, C. (1989). Response styles and ways of knowing. In C. Anson (Ed.), Writing and response ( pp. 332 – 365). Urbana, IL: NCTE. Bakhtin, M. M. (1981). M. Holquist (Ed.), The dialogic imagination. Austin: University of Texas Press (C. Emerson, M. Holquist, Trans.). Bates, L., Lane, J., & Lange, E. (1993). Writing clearly: Responding to ESL composition. Boston: Heinle and Heinle. Beebe, L., & Takahashi, T. (1989). Sociolinguistic variation in face-threatening speech acts. In M. Eisenstein (Ed.), The dynamic interchange: Empirical studies in second language variation ( pp. 199 – 218). New York: Plenum. Brannon, L., & Knoblauch, C. H. (1982). On students’ rights to their own texts: A model of teacher response. College Composition and Communication, 33, 157 – 166. Brophy, J. (1981). Teacher praise: A functional analysis. Review of Educational Research, 51, 5 – 32. Cardelle, M., & Corno, L. (1981). Effects on second language learning of variations in written feedback on homework assignments. TESOL Quarterly, 15 (3), 251 – 261. Cohen, A. D., & Cavalcanti, M. C. (1990). Feedback on compositions: Teacher and student verbal reports. In B. Kroll (Ed.), Second language writing: Research insights for the classroom ( pp. 155 – 177). Cambridge: Cambridge University Press. Connors, R. J., & Lunsford, A. (1993). Teachers’ rhetorical comments on student papers. College Composition and Communication, 44, 200 – 223. Daiker, D. (1989). Learning to praise. In C. Anson (Ed.), Writing and response ( pp. 103 – 113). Urbana, IL: NCTE. Dragga, S. (1986, March). Praiseworthy grading. A teacher’s alternative to editing error. Paper presented at the Conference on College Composition and Communication, New Orleans. Ferris, D. R. (1995). Student reactions to teacher response in multiple-draft composition classrooms. TESOL Quarterly, 29 (1), 33 – 53. Ferris, D. R. (1997). The influence of teacher commentary on student revision. TESOL Quarterly, 31, 315 – 339. Ferris, D. R., Pezone, S., Tade, C. R., & Tinti, S. (1997). Teacher commentary on student writing: Descriptions and implications. Journal of Second Language Writing, 6, 155 – 182. Gee, T. C. (1972). Students’ responses to teacher comments. Research in the Teaching of English, 6, 212 – 221. Hedgcock, J., & Lefkowitz, N. (1994). Feedback on feedback: Assessing learner receptivity to teacher response in L2 composing. Journal of Second Language Writing, 3 (2), 141 – 163. Herbert, R. (1990). Sex-based differences in compliment behaviour. Language in Society, 19, 201 – 224. Holmes, J. (1988). Doubt and certainty in ESL textbooks. Applied Linguistics, 91, 20 – 44. Holmes, J. (1995). Women, men and politeness. London: Longman. Hyland, F. (1998). The impact of teacher written feedback on individual writers. Journal of Second Language Writing, 7 (3), 255 – 286. Hyland, K. (1998). Hedging in scientific research articles. Amsterdam: John Benjamins. Hyland, K. (2000a). Disciplinary discourses: Social interactions in academic writing. London: Longman. Hyland, K. (2000b). Hedges, boosters and lexical invisibility: Noticing modifiers in academic texts. Language Awareness, 9 (4), 179 – 197. Johnson, D. M. (1992). Compliments and politeness in peer review texts. Applied Linguistics, 13, 52 – 71. Keh, C. (1990). Feedback in the writing process: A model and methods for implementation. ELT Journal, 44 (4), 294 – 305. Knoblauch, C. H., & Brannon, L. (1984). Rhetorical traditions and the teaching of writing. Upper Montclair, NJ: Boynton/Cook. 210 F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 Leki, I. (1990). Coaching from the margins: Issues in written response. In B. Kroll (Ed.), Second language writing: Research insights for the classroom ( pp. 57 – 68). Cambridge: Cambridge University Press. Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. Beverley Hills, CA: Sage. Myers, G. (1989). The pragmatics of politeness in scientific articles. Applied Linguistics, 10, 1 – 35. Onore, C. (1989). The student, the teacher, and the text: Negotiating meanings through response and revision. In: C. Anson (Ed.), Writing and response ( pp. 231 – 257). Urbana, IL: NCTE. Pennycook, A. (1996). Borrowing others’ words: Text, ownership, memory and plagiarism. TESOL Quarterly, 30 (2), 201 – 230. Reid, J. (1994). Responding to ESL students’ texts: The myths of appropriation. TESOL Quarterly, 28 (2), 273 – 294. Scollon, R. (1994). As a matter of fact: The changing ideology of authorship and responsibility in discourse. World Englishes, 13 (1), 333 – 346. Searle, D., & Dillon, D. (1980). The message of marking: Teacher written responses to student writing at intermediate grade levels. Research in the Teaching of English, 14 (3), 233 – 242. Severino, C. (1993). The sociopolitical implications of response to second language and second dialect writing. Journal of Second Language Writing, 2 (3), 181 – 201. Sperling, M. (1994). Constructing the perspective of teacher — as reader: A framework for studying response to writing. Research in the Teaching of English, 28, 175 – 203. Taylor, W. F., & Hoedt, K. C. (1966). The effect of praise upon quantity and quality of creative writing. Journal of Educational Research, 60, 80 – 83. Zamel, V. (1985). Responding to student writing. TESOL Quarterly, 19 (1), 79 – 101. Appendix A. Teacher interview prompts Attitudes to teaching writing 1. Could you describe the approach to teaching essay writing that you usually use when teaching ESL/EFL students? 2. What do you think is the biggest problem for EFL/ESL students when they try to write academic essays? Approach and attitudes to giving feedback 1. How helpful do you think teacher written feedback is for improving students’ writing? 2. What do you think is your main role when you respond in writing to a student’s draft? 3. What do you think is your main role when you respond to a student’s completed writing? 4. When you respond to drafts of student essays, are there any aspects of the texts which you focus on more than others? What are they? 5. Do you focus on the same aspects when responding to final drafts? 6. Do you think that teacher feedback is more helpful on a draft during the writing process or on completed writing? Why do you think so? 7. When you give feedback, which of your comments do you expect to be most useful to students to help them improve their writing? F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 211 8. Thinking specifically about language problems, what form of written feedback on language problems in drafts of writing do you usually give to students? Attitudes to other forms of feedback 1. Do you think teacher feedback given orally, i.e., in an individual conference or while walking around the class, is more or less helpful than written feedback or are they about as helpful as each other? Why? 2. How useful do you think peer feedback is to ESL students in helping them improve a draft of their writing? Why do you think so? Expectations of student behaviour 1. How do you expect students to use the written feedback you give them on their writing? 2. What would you expect students to do if they could not understand your comments or could not correct the mistakes in their writing after receiving feedback? Reflections on own experiences of feedback 1. Could you describe an occasion where you felt that you have given a very successful feedback? This might be to an individual or to a whole class. It might involve one episode or a treatment carried on over a whole course. 2. If you were to pass one piece of advice about giving effective feedback to a new teacher of ESL writing, what would that be? Appendix B. Student retrospective interview prompts General overview of the revisions carried out 1. How long did you spend revising this draft? 2. Could you describe what you did as you revised? For example, did you read the feedback first or did you refer to the feedback as you revised? 3. What were the main changes you made to the draft? 4. What do you think was the most important change you made to the draft? General overview of feedback use 1. What was the most useful feedback your teacher gave you on this draft? 2. Did you get feedback from any other source? 3. What use did you make of your peers’ comments? On global comments and changes 1. Why do you think your teacher made this comment? 212 F. Hyland, K. Hyland / Journal of Second Language Writing 10 (2001) 185–212 2. What changes did you make to the writing after you read this comment? 3. Do you feel more satisfied with your writing now? 4. Why (not)? On localised comments and changes 1. What do you think this comment is asking you to do? 2. What change did you make to your writing because of this comment? 3. Do you think your change has improved the writing? 4. How has it improved your writing? On comments and corrections ignored 1. Why do you think your teacher made this comment? 2. Why did you not make any changes to the writing? 3. Do you think there is still a problem with the writing? Student evaluation of their success in revising 1. When you look at your first and second drafts, do you feel satisfied with your revisions? 2. Do you feel that the essay has improved? How? 3. Is there anything about writing that you learned from writing this essay that you will remember and use in the future?