Error Correction: a Critique

For Initial Teacher Training qualifications error correction is a major topic. Both the Cambridge CELTA and the Trinity Certificate in TESOL contain criteria on dealing with errors. They encourage trainees to deal with errors in a variety of ways, such as the following, which have been taken from Learning Teaching by Scrivener (2011: 285 – 290):

  • Indicating an error has been made
  • Eliciting the correction through questions
  • Finger correction
  • Echoing
  • Opening the error up to the rest of the class

Although there is quite an extensive list of suggestions and techniques, they essentially boil down to two approaches for dealing with error:

  1. Immediate Error Correction
  2. Delayed Error Correction

Immediate Error Correction

The vast majority of ideas found in methodology books for dealing with errors focus primarily on the first approach: Immediate Error Correction. They spell out in detail error correction techniques for teachers because the most natural response for the teacher would be to simply correct the learners. This is seen as poor practice because it probably wouldn’t lead to anything conducive, as “the more the students are involved in correction, the more they have to think about the language…” (Edge 1989: 27). So it seems by giving teachers very clear learner-centred techniques for dealing with errors, learners will become more “aware” and “notice” their errors (Scrivener 2011: 285).

Delayed Error Correction

The second approach also takes a learner-focused stance, in that it encourages the learners to “discover” for themselves where they have gone wrong, with the hope their greater cognitive engagement will lead to them “remembering” the language better (Scrivener 2011: 285). Although there are a variety of ways to do this, their basic premise is the same: to collect samples of the learners’ errors to later feed back to the learners as an exercise.

Delayed Error Correction is probably most often conducted as a short whiteboard activity, where the teacher writes up a handful of incorrect sentences (or possibly a mixture of both correct and incorrect) and the learners have to identify where the issues are.

Issues with Delayed Error Correction

The idea behind this approach is simple yet beautiful: bring together a sample of errors and get the learners to tell you what they have done wrong. The only problem is that as each teaching year goes by, the more it seems to me this technique does not achieve the desired effect.

Firstly, unless the error was a slip of the tongue, the learner who made the error probably won’t know it is an error. This leaves a stronger learner to correct it. What does this achieve? Not much: the stronger learner has corrected something which they probably don’t make errors in; the weaker learner has failed to engage with the language in such a cognitive way that they will continue to make the same error.

Secondly, a lot of errors don’t make sense out of context – as does a lot of correct language! Take this sentence as an example, “You took the chicken out?” If this were dealt with as an individual item, it would be flagged up as an incorrectly formed question. However, in the context of a mother asking her son if he had indeed remembered to do what she had told him to do some time earlier i.e. take the chicken out of the freezer, then it makes perfect sense and is even stylistically and grammatically correct. So, putting on the whiteboard ‘He never play rugby’, what is it the learner is supposed to be correcting: He never plays? He has never played? He will never play again? Without a clear context, this attempt is a clear waste of time. Given this is delayed error correction, the context could have well been significantly earlier in the lesson.

Finally, as you go down the CEFR scale, the more and more learners think what you say, write and do is of the utmost importance. This ultimately leads to a situation which I have seen time and time again: listening in to the learners during a task, the teacher boards samples of their errors, and as they do this, the learners stop their task to focus their attention on the ‘important’ thing the teacher is writing on the whiteboard. This is when learners see the teacher as a ‘knower’ and more often than not, it leads to a situation which the whole approach is supposed to avoid: learners begin to immediately attempt to correct the errors. What’s more, the moment one of them starts shouting out possible corrections, the rest stop what they’re doing to join in.

Why Bother?

Thornbury (Online) says that many perceived errors, such as not adding the third person –s morpheme, are “an inevitable stage of language learning” and they are “extremely resistant to correction.” Furthermore, errors do not seem to be a matter of L1 interference – they are predictable at each stage of learning the language. They must, therefore, be an integral part of language acquisition and remain unchanged until the learner has acquired that feature of language.

All of this suggests error correction should be dropped altogether. However, that might have a negative effect on the learners’ expectations: learners want and expect error correction– they might also feel “short-changed” if they don’t receive any teacher-led corrective feedback (Pinard Online).

The Solution

If error correction has to be present due to learner expectation, yet it remains largely ineffective in preventing errors, then it seems the most obvious compromise is to correct errors at the moment where they most directly impede performance (both task performance and linguistic performance).

Most lessons will have a practice stage – if you’re delivering a PPP lesson, then this stage is the second P; for TTT lessons, this will be part of the second stage – it is during this stage where error correction might be most useful for the following reasons:

  • Learners are concentrated
  • Learners are focused on the target language
  • Learners are using their existing knowledge to test out and ‘play with’ the target language
  • Learners will expect 100% teacher-led correction at this stage
  • Learners are cognitively aware and honed in on the target language

During this stage there is a greater focus on accuracy. Methodologists, such as Scrivener (2011) and Harmer (1989), identify accuracy-focused stages as moments where error correction could be most effective. When the lessons moves on to the final production stage (or second task stage) it really is necessary for the teacher to take a step back and let the learners get on with the task at hand: whatever mistakes they make will simply be inevitable and an integral part of the learning process.


Edge, J. (1989) Mistakes and Correction. London: Longman.

Pinard, L. (Online) DELTA Notes 1: Error Correction. Available at:

Harmer, J. (1989) The Practice of English Language Teaching. London: Longman.

Scrivener, J. (2011) Learning Teaching: The Essential Guide to English Language Teaching. 3rd Ed. Oxford: MacMillan Education.

 Thornbury, S. (Online) Error Correction. Available at:

6 thoughts on “Error Correction: a Critique

  1. You brought up some great points on error correction. I have never been a big fan of delayed error correction for many of the reasons you discussed above but mainly because of the context factor. I’ve noticed that a correction ‘in the heat of the moment’ seems more effective and memorable for students. Having said that, I also believe that there are ways to use delayed error correction, such as the strategy outlined above by Hada. In my YL classes I use delayed correction when we’re giving feedback on tests. I put the most common mistakes on the board and I get them to discuss what the correction answer would be. As Hada mentioned, I only let them discuss the corrections and not write them because I think it would be distracting.
    Thanks for such a thought-provoking post. It’s nice to hear other teachers’ perspectives on techniques such as error correction.


    1. Thank you so much for commenting! I really appreciate it when someone takes the time to comment – I guess it also shows me my stuff is being read 🙂

      Apologies for the huge delay in responding: in the process of catching up now with everything after a very busy period.

      Perhaps somewhat ironically, after writing this post, I gave a class a test, and then used the issues in the test for the subsequent two lessons. It meant largely revisiting what we had already done but the test performance really strongly influenced the content of those two lesson. Bit like your YL classes you mentioned.

      During one of those lessons, I had an interesting moment quite related to what you said about a hot correction being more ‘effective’ – the learners were making subject and object questions, and one of them was basically asking me for the answer by offering up a variety of possibilities. My training was like “don’t give away the answer, they have to engage with the material!” but my gut insight was like “look, if you say which one is correct, that could just be the in-the-spur-of-the-moment input they need to get this clear in their heads.”

      Sometimes, I wonder if we’d be better off trained in psychology than pedagogy….


  2. Thanks for this post – a good summary of the main theories on error correction.
    I often use the delayed approach so I’m not sure I fully agree with you that it ‘does not achieve the desired effect’ although I must admit that I apply it in quite a systematic way.
    1) First, I often tend to use this approach at the beginning of lessons, and I find that over the course of a 6 week term, students pick up on some fossilised errors as a result of the repeated errors they are exposed to.
    2) I always make sure I ‘recontextualise’ the sentences. So for example, if I begin the lesson by asking them, ‘what did you do yesterday?’ with the objective of checking their use of the past simple and building their lexis on everyday chit chat, I’ll only highlight the errors connected to that.
    3) Many students comment that they like this approach as it gives them a chance to be exposed to the same errors several times over the course of the term and this has a noticeable impact on their confidence which in turn works on both their fluency and accuracy..
    4) I make sure they work out the mistakes in groups and insist they don’t write anything during the activity. They usually take a picture of the board at the end and share it on the class WhatsApp group which gives the weaker students a chance to review what was said in class.
    All in all, I think that like many of the approaches we find in methodology books, tweaking and adapting is often key to make them effective.
    Thanks again for a great post.

    Liked by 1 person

    1. Sorry it has taken me so long to respond – been terribly busy!

      Thank you so much for commenting – I really appreciate it when people comment: I wish more did!

      It’s been a while since I was in a more regular teaching post, where I see the same learners every week, but I’ll get back to that around March 2015.

      When I get back to regular teaching, I want to make use of what you’ve highlighted. I think delayed error correction in this way – spanning several lessons, being used for worksheets, revision etc – could be really effective for dealing with fossilisation. Certainly a colleague of mine at IH Torun, Richard Lacy, recommends it for that.

      With regards to your last point: by not letting them write anything down, they then have to talk about the errors with their partners – is my understanding correct?

      Liked by 1 person

Tell us what you think...

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s