As I read Hugh Moss's article "Reflections on Written Feedback" in issue N" 10 of The Journal, I was struck by the number of echoes I heard of my own experience. In this article I would like to take some of the points raised there further and add a few new ones.

For the learner to benefit fully from a writing task, especially if it is a 'full-length' piece of writing such as a letter, report or composition, a lengthy process is involved. To begin with, the task has to be properly prepared, normally in class. This preparation may involve work on the specific grammar and vocabulary it demands. It will almost certainly include work on how the writing is to be structured. It is then our hope that when the student comes to the actual writing stage, he or she will take due care to incorporate what we have taught. Certainly, the conscientious student will try to do so.

And if we are going to go to all this trouble to prepare the learner, who in turn goes to the trouble of taking care at the writing stage, the feedback on this writing must be equally conscientious. Otherwise the whole exercise will have been more or less a waste of time.

To emphasise thorough feedback is particularly important in a context such as we find all too often in Portugal - that, in the students' minds at least, marks count for everything and real learning for not very much. Indeed, we can only counter this mentality if we refuse to let students merely put their work away when we have returned it and they have seen the mark, by continuing the learning process through a careful analysis of the feedback.

Like Hugh Moss, I have tried using symbols for students then to correct their own mistakes. Though I have had colleagues who tell me that this works well, I have to confess I've always been disappointed with the result. Again, like Hugh Moss, I have come to the conclusion that this is inevitable - if the students could correct themselves, then in most instances they presumably wouldn't have made the mistakes in the first place.

I have also agonised over just how fussy to be. I remember that when I was starting my career there was a certain fashion, which recurs cyclically, for not correcting everything. There were two main arguments put forward for this: firstly, it means you can focus better on the really important errors; secondly, it avoids the disheartening appearance of a writing task returned covered with red or, slightly more kindly, green ink. While I have some sympathy with both points, I have never wanted to put this policy into practice. Furthermore, repeated comments from students over the years have confirmed my view that work needs to be fully corrected. This is what they expect, feeling sold short otherwise and preferring honesty to 'kindness'.

What we are left with, then, is no symbols and a script laden with teacher ink - ideally not just corrections, but, again as Hugh Moss points out, with comments, annotations, maybe a summary of main points at the end. A very time-consuming task that deserves a better reward than just looking at the mark and filing.

So what are we to do?

Over the past few years I have developed two distinct but complementary strategies: one which focuses on the gravity of error and one which looks at some of the psychological processes that lie behind it.

Gravity of error

One of the virtues of the don't-correct-everything movement is this notion that some errors are more serious than others. Individual teachers might not agree on what constitutes a grave error and what doesn't and every teacher even of quite limited experience could rattle off a list of errors they particularly love to hate. At higher levels there comes a point when certain types of error become simply 'unacceptable', particularly in contexts where accuracy matters, such as public examinations and the work of language professionals, practising and future, like teachers and translators.

But as soon as we use a term like 'grave' or 'unacceptable' we need criteria by which to define them. Otherwise what for one teacher/marker may be 'unacceptable' another might not regard as particularly 'grave'. In the two universities where I teach, I have increasingly felt the need for a 'list', a piece of paper that I can refer to so that I have something concrete to show a student. That way I can tell him or her, "This is why the piece of work just isn't good enough. Look how many of your errors are on the list." But what should be on the list'?

I decided that 'the list' needed to be in some sense objective. It could not be based simply on my own love-to-hates. After all, while teachers may agree on quite a lot, they also develop idiosyncratic irritations among student errors. Similarly, to gather a group of colleagues together, while it might help to identify a common core, it would also lead to a multiplication of the 'idiosyncratic irritations'. I decided, therefore, to build a list based on an existing document and therefore not subject to the hates of individual teachers: the Ministry of Education's official syllabuses. I felt that nobody could reasonably object to an insistence that everything from the 1st two years of English, the English taught in the 2º Ciclo, should be fully mastered by university level. Many indeed would, argue that this is too low a threshold but the fact of the matter is that many university level students continue not to have mastered even these basic structures. The list (now in a second revised version) looks like this:

I also felt that by university level, even in the first two years, there were many other basic things that should have been fully mastered, but which were missing from this list in large part because they are only taught in the third and fourth years of English (7º/8º anos). This Ied me to compile a second list: 'Other grave errors'.

Now it is one thing to have a list but what then do you do with it? In the first place the students need to be fully aware of the notions of 'grave' and 'unacceptable' errors. In the case of the unacceptable ones, they also need to know the basis for their being unacceptable, namely that they refer to items that were taught in the first two years of English. Secondly, they need a copy of the list, so that they can refer to them. As a teacher, you need to go through the list making sure your students understand the points at issue, if necessary explaining. The list can also serve as a basis for remedial work and spot-the-mistake exercises.

However, coming back to the correction of written work, I have adopted a system of marking two crosses above corrections to 'unacceptable' errors and one cross above 'other grave errors'. In this way the student can quickly see both how many of each kind there are and where. This enables the student to focus immediately on the most serious errors. This is already useful in itself, but to help the students gain the full benefit from this the teacher can then set the task of consulting the list and identifying which item on the list is involved in each instance. This then tells the student what in particular they need to rectify in order to reduce and eventually eliminate their more serious errors. In this way, your feedback is not just filed but clarifies to the student areas for further work.

A side benefit of indicating these errors with crosses is that the number of unacceptable and other grave errors is also immediately obvious to the teacher. I find this actually helps me in giving the work a grade.

Psychological processes

Learners need not only to be aware of how serious their errors are and the nature of their more horrendous transgressions but what led them to error in the first place. For this reason I devised a second strategy for following up written feedback.

The detailed reasoning that leads to error is exceedingly complex. The careful, reflecting student may indeed go through a whole series of trains of thought in agonising over a single word or phrase. While it would be unrealistic to try to analyse the detailed processes that lie behind each word that needs correcting (even if the learner could remember them all), we can, I believe, make a useful threefold division.

Firstly, there are 'silly mistakes', careless slips, the ones that make you feel like shooting yourself through the head when they are pointed out. These slips have two important characteristics: the student instantly recognises them as errors, being able to correct them unaided, and, by definition, they do not recur. In my view, they are the only true errors. Anything else that needs correcting is not the consequence of making a mistake, for there was never a great possibility of getting it right. It is rather the product of trying to express something for which the learner's current knowledge is insufficient. It is the symptom of a 'problem' that the learner cannot solve alone, only with the help of someone who does know how to express it.

The second area, then, is a specific type of 'problem': the translation problem. Translation problems can affect all aspects of the language, not only such obvious areas as lexis and grammar, but also spelling and punctuation. Furthermore, while it is most often the learner's mother tongue that lies behind this type of problem, the interference may be from some other language. (Among my students of English and German, for example, I regularly find interference from German in their English.) Up to a certain level, this type of problem is not only common but extremely natural. It is also characteristic of the truly advance learner that they have largely or completely overcome this problem.

Thirdly, though not very scientifically, are 'other problems': the words and expressions that need correcting but cannot be explained away simply as slips or translation problems. In practical terms, I find that these can usefully be subdivided into two: those that the learner grasps easily once the correction has been made and those that continue to puzzle the learner and therefore need not just correction but explanation.

To exploit the threefold silly mistakes/translation problems/other problems distinction, I use the following procedure:

1. Tell your students to go through their work and number 'your corrections' from 1 to however many there are. (Note, I do not refer to 'their mistakes' but 'your corrections' - they are, of course, the same things, but it prepares for the distinction between 'mistakes' and 'problems' I made above.) Stress that every correction must be numbered, including spelling, punctuation, etc.
2. On the board, write up these categories for your students to copy down: Silly mistakes Translation problems Other problems: you understand the correction Other problems: you don't understand the correction
3. Explain what is meant by each of these. Then instruct your students to go through the corrections and according to which type it is, write each number beside one of the categories. Stress that 'translation problems' affect not only grammar and vocabulary, but also spelling and punctuation - anything that transfers the practice of one language onto the one they are currently working on.
4. Go round the class, gathering the translation problems and writing them on the board. I usually do this by writing, first, what the student wrote, then, in brackets, the word/expression that has been translated, and, finally, what it should be, eliciting this from the class, e.g.

in what concerns (no que diz respeito) = as regards
Talk through the problem and the correction as you write it up.

5. Ask your students which corrections they put in the final category. If they don't understand the correction, you need to explain it.

I have found, by experience, that the translation problems could, in many instances, just as easily have been made by most of the other members of the class and therefore going through them like this forms a valuable input session, often involving small points that a teacher would never build a whole lesson around. The reasons for explaining the corrections that the students don't understand are obvious. Silly mistakes do not need to be followed up and the 'other problems' that the student does understand do not need follow up. To do so after going through translation problems tends to lead to an excessively long session, which becomes boring and hence, to a large extent, counterproductive. If, however, I were working in a multi-lingual context, it would be the translation problems I would omit and the other problems, in general, that I would go through with the class.

There are two small points I would wish to add, by way of comments on this procedure. Firstly, students do not always correctly identify what category of correction is involved in each case. In particular, they often dismiss as 'silly mistakes' corrections that indicate a more deep-seated problem. If the same 'silly mistakes' recur in subsequent pieces of work, this indicates clearly that there is a more systematic problem involved. Secondly, this procedure helps to show students how serious a problem 'translation' is for them, an aspect that varies considerably from student to student and which learners are not always sufficiently aware of.

Conclusion

I would stress the importance of going through corrected scripts in this way. I feel there is such a lot to be gained from doing so. Both of the procedures I have described, in different ways provide useful follow-up. I generally use one or other with any particular piece of written work, but they can be used together, especially when your students are used to them and can therefore do the tasks more speedily.