Giving Learners a Voice in Correction and Feedback

by Peter Watkins

University of Portsmouth (Portsmouth, UK)

Keywords: agency, correction, feedback, learner autonomy



This paper reports on one attempt to give learners a greater degree of control of the correction and feedback they receive on their own language production. It begins by establishing the usefulness of correction but argues that a broader conceptualisation of feedback on learner production is required. This conceptualisation moves the teacher’s role away from being judgmental and towards being supportive and collaborative and from this more equal relationship between learner and teacher comes a greater voice for learners in shaping the feedback they receive. A simple, small-scale experiment is described and the results suggest that it may be possible to train learners to be able to ask questions about their own and other’s output and in so doing, shape the feedback they receive.



When people learn an additional language in a classroom context, it is distinguished from first language acquisition in a number of ways. One major difference is the role that correction plays, as Cook (2001, pp. 182-3) notes:

 Negative evidence by correction is … different in L2 learning. In the first language, it  is not so much that it is ineffective as that it occurs rarely. In the second language,  correction of student errors can, and often does, occur with high frequency. The L2  learner has an additional source of evidence not available to the L1 learner.

So, dealing with errors is an integral part of an L2 teacher’s job and is therefore worth investigating. However, it characterises teacher feedback on learner language production in a limited and largely negative manner – only applying when things go wrong with that output produced by learners. This paper will look briefly at traditional views of error correction, but will then go on to look at the arguments for taking a broader perspective on giving feedback to learners, in which correction plays an important part but is not the only element of the process. This broader view results in feedback routines being more supportive and less judgmental in nature than is the case with traditional error correction. In this environment learners are better placed to take on more responsibility for the feedback process and the paper will conclude by reporting on one small scale experiment that embedded this in a classroom procedure. This paper focuses on oral language production, but the arguments could apply equally to written language.


Dealing with error                                                                                     

Error is a problematic term because it implies a deviation from a norm, which leaves us with the problem of defining what that norm should be. It is not within the scope of this paper to enter that debate although it is worth remembering that native speaker norms (themselves difficult to define) are no longer axiomatically regarded as the gold standard by which learner production should be judged. In addition, it should be remembered that the appropriacy of a piece of language is affected crucially by the context in which it is used, thus increasing the difficulty of making some of those ‘right’/‘wrong’ judgments.

Despite these inherent difficulties, most teachers intuitively see the benefit of appropriate error correction and there is some research evidence to support this position. Ellis (1997, p. 54) cites research by Lightbown and Spada who found some evidence to support the view that correcting errors when they occurred spontaneously contributed to the acquisition of some (although not all) target language forms.

Of course, saying that error correction is likely to be useful is very different from saying how it can most usefully be achieved. Larsen-Freeman (2001, p. 40) notes that

While most teachers value using feedback to help students bring their interlanguage into alignment with the target language, questions of how much and what sort of feedback to give students on their grammatical production are unresolved.

Without the benefit of clear research evidence to guide practice, teachers use their common sense and base their decisions on experience and intuition. The pedagogical issues that teachers typically deal with are summed up by Scrivener (2005, p. 299):

  1. What kind of error has been made (grammatical? pronunciation? etc.).
  2. Whether to deal with it (is it useful to correct it?).
  3. When to deal with it (now? end of activity? later?).
  4. Who will correct (teacher? student self-correction? other students?).
  5. Which technique to use to indicate that an error has occurred or to enable correction.

The received wisdom offered to teachers to help them answer these questions usually includes reference to the degree to which the error compromises communication, as clearly intervention is required if communication has broken down but is optional if it hasn’t. Also, if the focus of the activity is accuracy-based then more immediate correction may be favoured, but in fluency activities a teacher may be advised to delay correction work so as not to interfere with the communication flow. Whether or not the language item is already known to the learners is another variable, and teachers may also make decisions based on individual characteristics of a particular learner, such as the degree of nervousness they experience when speaking in front of a group.

All of these considerations seem sensible and useful for teachers to keep in mind, and it is reasonable to assume that as teachers gain experience, so they become highly competent in both recognising where correction is required and also adopting appropriate strategies to correct learners.

Despite its obvious worth, error correction poses some difficulties from a methodological perspective. It implies a vision of learner output that is in deficit to some external standard, placing that output as something to be judged by the teacher. To some extent this is at odds with ideas of the teacher being a classroom facilitator who can support learner production and also with notions of developing learner autonomy as the amount and timing of correction is largely outside a learner’s control. In order to address these issues we will look at teacher feedback on learner output from a broader perspective.


Feedback: A broader perspective

The first issue here is partly about terminology. Correction implies that something has gone wrong and that it has to be put right. Typically, although not exclusively, teachers tend to focus on problems of form, relating to grammar, lexis and pronunciation. The term feedback can have connotations of much wider reference, making it appropriate for a communicative classroom, where it is reasonable to expect to see feedback on what is said (the communication element), rather than just how it is said. For example:

 Learner: My daughter passes her driving test yesterday.

 Teacher: That’s wonderful – did she pass first time?

The communication driven response of the teacher (rather than a form-driven response which would focus on passes) encourages the learner to say more, and by giving the learner further opportunities to interact meaningfully, the teacher can maximise learning opportunities in the classroom. It should be noted that there is no reason why this kind of content-based feedback cannot be combined with a form focus, which may come at a later stage.

Another type of meaning-based feedback is that which occurs in our every day speech, the so called ‘back-channelling’ that listeners provide to speakers to show that they are engaged with, and following, what is being said. A key point to notice in both these examples is that the teacher’s contribution is to support (or ‘scaffold’) the communication and to facilitate its development. According to socio-cultural learning theory (Vygotsky, 1978, for example), the learner is able to achieve more than would otherwise be the case by interacting with a ‘more skilled other,’ the role fulfilled by the teacher in the example above, although it could equally be provided by other learners. It is this scaffolded interaction that drives language acquisition.

In addition, by providing feedback in these ways the teacher is highlighting that the communication is being successful (without necessarily being perfect) and this in itself is likely to be motivating to the learners. There is also the added advantage that by using these features of naturally occurring speech, the teacher gives all the learners in the class the opportunity to observe, and participate in, authentic discourse management, an essential skill to develop if learners are to go on to be successful communicators.

The examples of feedback given above occur in everyday speech, but feedback can also have a clear and specific linguistic focus. Whereas error correction focuses on what went wrong in the learners’ production, feedback can additionally focus on what went right. For example, when a teacher conducts feedback after a group speaking activity designed to develop fluency (often the time that correction is provided) the teacher can also highlight good examples of language use, or simply features of language which seem useful to learners. This may help other learners in the group to ‘notice’ the features of language that separate their language from that of more expert users and so may aid the acquisition of those features in some circumstances. So important is this teacher role of responding to and shaping output that Meddings and Thornbury (2009, p. 8) comment that

 The teacher’s primary function…is to optimise language learning affordances, by, for example, directing attention to features of the emergent language.

The view of the teacher’s role expressed here is consistent with ‘deep end’ views of communicative language teaching, in which the teacher sets up the conditions in which communication can take place and then helps learners to notice how their own language is different to that of more expert users. However, the role could be extended beyond the teacher directing attention to salient features of language to include the training of learners to be able to ask questions about features of language in the texts they themselves have created. This is not a new idea and is a cornerstone of Community Language Learning, for example, where a text generated by learners is frequently transcribed and analysed, with the analysis being driven by learners’ comments and questions (see Richards & Rodgers, 2001, pp. 90-99, for a fuller description).

In the next section of the paper we will move on to look at arguments for giving learners greater influence over the type of feedback they receive.


Giving learners agency

One feature of correction is that it tends to be teacher-controlled. Although many teachers may encourage a learner to self correct, or indeed invite other learners to suggest alternatives, the process of correction is usually instigated by the teacher. More learner-centred approaches to teaching have not necessarily had a great impact in this area. As we noted above, a teacher’s view of when and how to correct may be influenced by many things such as the type and stage of the lesson, their own experience as a language learner, or their own preferred learning style. Of course, a teacher will probably vary both the amount of correction that is offered and also the method of correction to suit the circumstances of the class, but the fact remains that in most classrooms the teacher remains in control of correction.

However, it is possible to give learners a greater degree of agency in this area. Byagency I mean control of their own learning. Thornbury (2006, p. 10) claims that this can have long term benefits to the learner:

 A ‘sense of agency’ has also been identified as a factor that contributes to  motivation. Learners who feel they are responsible for their own actions are likely  to go further.

But in a practical sense how can this ‘sense of agency’ be achieved when it comes to giving feedback? As is the case with Community Language Learning, one way is to analyse tapescripts of earlier production, but this can sometimes lead to quite long time delays between the actual production when the language was meaningful to the learner and the analysis taking place. The next section of the paper looks at one small-scale experiment that considered whether learners could be trained to take more control of the feedback process more immediately.


Achieving learner agency in feedback

This section will describe a small scale experiment designed to increase learners’ awareness of their own language production and their ability to ask questions about it, thus giving them a degree of agency over the feedback received. Two business English classes at approximately B2 level on the Common European Framework of Reference were used. One (Group A) used the procedure described below. The other (Group B) was used as a control.

In the first phase of the experiment, both groups used pair work activities based around moral dilemma situations in a business context. So, for example, learners discussed what they would do if they learned that their boss was over-claiming expenses for business trips. In Group A, however, each pair was shadowed by another pair (creating a group of four) with each observer watching a particular participant and making notes about their language output. At the conclusion of the pair work activity, the observer passed the notes onto the person they observed, who could either choose to ignore the notes, or ask for guidance from the ‘observer,’ another peer or the teacher. It should be noted that the decision of whether to ask, and who to ask, for clarification remained with the participating learner, thus giving them a sense of control over the process. My hope was that this activity would lead to greater awareness on the part of the learner and would help them to formulate their own questions about language in the future. Group B was given standard teacher instigated feedback on aspects of language output that I heard as I monitored the pairs.

Before Group A started the pair work/observation task, we reviewed question forms that may be useful in analysing texts. For example: Is this section clear? Is there a more (in)formal way of saying…? How would the meaning change if I used …? Would it be wrong to say x in that example? and so on. We then analysed a short transcription made from an earlier lesson, giving the opportunity to use the forms above. This lead in was designed to give learners the tools they needed to ask about language and also to support the observers in their task as it helped to indicate the areas that they might comment on. A discussion of content and the ideas that were expressed was conducted separately before the more form-focused issues that sprang from the notes supplied by the observers.

As one would expect, after the pair work activity, there were far more questions from Group A about the language they had produced than from Group B. Most were addressed to the teacher, although several queries were resolved within groups.

Two weeks later both Groups A and B did a short role play based on the location of a new supermarket. On this occasion both groups were given the same task preparation, simply a short discussion about supermarkets and shopping. After the role play I had planned to ask if the learners had any questions about the language they had used, or found they needed during the role play. However, this prompt was not required in Group A, where, after the role play, the 17 students spontaneously produced nine questions relating to language they had used, or tried to use. There had also been several questions asked as I monitored the groups. The 19 students in Group B asked only three questions between them and were prompted to do this (see above).

Of course, inevitably with a small scale experiment, the results could be explained in various ways, including that enough learners in Group A had preferred learning styles that predisposed them to asking such questions and thus skewed the results. In order to test this explanation I have since used the treatment outlined above with the learners in Group B. The end result was again that learners asked more questions (eight) without any prompting from the teacher. Although, of course, it may be that the questions of one learner act as a prompt to others.

It is hard to say definitively but it is possible that the learners have been trained to play a slightly different role during and after productive phases of a lesson, using questions to direct the teacher’s attention to the output on which they want clarification, instead of waiting for the teacher to determine all feedback. It would be interesting to monitor whether groups trained in the way described continue asking questions about their output over a greater period of time and also the degree to which the questions reflect their own perceived learning needs. Although not conclusive, the experiment may indicate that learners can be trained to analyse their own as well as each other’s output and to request feedback on it. By creating opportunities for this to happen the teacher is helping to give agency to learners in this area of language learning, as the learners develop the skills to request feedback when they want it.



Both research evidence and teacher’s experience suggests that error correction is an important part of L2 learning in a classroom environment. There is also good reason to believe that a broader understanding of feedback, which includes but is not limited to correction, may be useful. Traditionally, decisions about the amount, method and timing of feedback have been in the sole domain of the teacher but some preliminary research suggests that teachers may be able to empower learners to have a greater voice in these decisions.



Cook, V. (2001). Second language learning and language teaching (3rd ed.). London: Arnold. Ellis, R. (1997). SLA research and language teaching. Oxford: Oxford University Press. Larsen-Freeman, D. (2001). Grammar. In R. Carter & D. Nunan (Eds.), The Cambridge guide to TESOL. Cambridge: Cambridge University Press.

Meddings, L., & Thornbury, S. (2009). Teaching unplugged. Peaslake: Delta Publishing.

Richards, J.C., & Rodgers, T. (2001). Approaches and methods in language teaching (2nd ed.). Cambridge: Cambridge University Press.

Scrivener, J. (2005). Learning teaching. Oxford: Macmillan.

Thornbury, S. (2006). An A-Z of ELT. Oxford: Macmillan.

Vygotsky, L. (1978). Mind in society. Cambridge, Massachusetts: Harvard University Press.


About the Author

Peter Watkins is a Senior Lecturer at the University of Portsmouth, where he has worked extensively on both pre-service and in-service teacher education programmes. In addition to a number of articles, he has written Learning to Teach English (Delta Publishing, 2005) and is the co-author (with Scott Thornbury) of The CELTA Course Trainee Book and The CELTA Course Trainer’s Manual (Cambridge University Press, 2007).

Leave a Reply