Monday, 13 April 2015

How much should language teachers correct errors?

I've been doing some superficial reading on the value of error correction in the classroom following a Twitter discussion with Chris Stolz, an American TPRS teacher who strongly supports the view, advanced by Stephen Krashen back in the early 1980s, that error correction plays no part in improving students' language proficiency.

Like most teachers I used error correction of writing, sometimes selective, sometimes just underlining with or without codes in the margin, along with selective correction and "recasting" of oral responses. I confess that I rarely questioned this general approach and assumed it to be appropriate. I and my colleagues would, however, on occasion discuss different types of written error correction. In reality, most correction we did was of written work. Ah... Good old marking!

To keep things brief, and because this blog is not an academic journal and you do not have much time, this is what I have found in my reading.

Krashen's comprehension hypothesis (all you need to acquire language is meaningful input with no focus on form or accuracy) gained currency from the 1980s and still has support. Teachers who believe in the sole use of comprehensible input and little or no correction can cite research to support their approach.

Since that time, however, a good number of other academic studies suggest that in the classroom setting traditional error correction and a focus on grammatical form do lead to an improvement in proficiency. Whether this is long term improvement is open to some doubt, partly because it is so hard to do reliable long term ("longitudinal") studies.

Research, therefore, is somewhat ambiguous and unhelpful, so far, with regard to error correction. You can pick and choose which academic to believe. I have the impression that the balance of research favours error correction of writing, which would confirm most teachers' gut feeling. As regards speaking, I see no reason to change my view that selective correction and recasting of responses is wise. Even those who support a fully naturalistic approach (second language learning is like child language acquisition) would have to acknowledge that parents correct their children and probably do so with good reason.

That said, there is plenty of evidence that large amounts of exposure to "comprehensible input" remains the key way to get students to acquire a second language. An obsession with accuracy and error correction is not the way to go.

For a good, balanced academic summary of this issue:

- Posted using BlogPress from my iPad


  1. Thanks for offering both sides to this issue. I'm sure there is a difference between error correction and effective error correction. I know that I valued both oral and written correction as a student of French and that it was effective because I acknowledged it and used it to improve. Error correction could be considered useless if the student did not recognize its importance or did not know how to apply it to their own language use.

  2. Thank you for leaving a comment, Dana. When research is unclear you can only go on your own experience and feeling. Some one tweeted today that research is their Satnav (GPS) but they use their own knowledge to get to the final destination!