Answers to the last post's questions about correction in language teaching

NB

1. In this and the previous post, a mistake or error is language that would be considered unsuccessful or inappropriate usage in the relevant social context. Even by this criterion, we often disagree about whether something is an error, and even more often about how serious an error it is. These notions are tied up with our understanding of the context in which the error was made.
2. The teaching contexts I have in mind involve adults only.

Should minor errors ever be corrected? (I mean things like wrong prepositions, word order in questions, missing articles, etc.) If so, why?
The thought underlying this question is: do minor errors matter? Should people care about them? The answer is that it depends on what you want to use the language for, and on your problem situation more generally. Some language learners want to be able to speak accurately as well as clearly. Others don’t. Both preferences are legitimate. For teachers, the context is obviously relevant here: if you are teaching exam preparation, it is more likely that learners will want to know about how the accuracy of their language could affect their performance.

So to answer the original question: only if the learner seems interested in being corrected on those things, and is developmentally ready for them. Observe how they react and adjust your corrections as needed. At higher levels you can simply ask: “Would you like some feedback?” and gauge the effusiveness of their reaction. If you still aren't sure, the safest thing is to use interactional recasts: find a way to echo their remark, incorporating the correction, while maintaining the flow of conversation.* People who are interested in being corrected tend to pick up on this, while those who aren't tend to ignore it – so it is both a way of correcting students and gauging their interest.

Another rule of thumb is: don't correct slips. These are mistakes that the learner makes in one instance (due to tiredness, echoing another learner, the complexity of his or her surrounding talk, etc.), but which you know they wouldn't normally make. All correction does in this case is draw attention to their lack of control in the L2 – likely not something they need you to point out.

How much context is needed for correction of an error to be meaningful or beneficial?
I was thinking of delayed correction here. There are many types of small error for which context isn't as important. Misunderstanding is always possible, though, so ideally the delay is short, and ideally learners would remember the context based on a phrase or a sentence.

Is it ever harmful to correct language learners' mistakes?
Yes, if it humiliates, annoys, frustrates, discourages, bores or distracts them. Pre-course needs analyses with a question about correction should be common practice for this sort of reason, although that is no replacement for teacher sensitivity.

In addition, some types of correction can be problematic when the learner's intended meaning is not clear. For example:


In this exchange, student J immediately takes up the teacher's recast – an apparently successful instance of uptake following correction. But it's possible that he/she was originally trying to say more than or less than. If there is any doubt about what the person means, it is better to seek clarification than to elicit a well-formed utterance that he/she may never have wanted to say in the first place.**

How often do learners even know what you're correcting them on? Is it worth checking? And if so, do you need to check every time, or just in specific cases?  
Not only do they often not know, they are often outright mistaken about it. You might correct a grammatical error and they will think it was a lexical error, or vice versa. They may even mistake a lexical correction as a (rather unfair) attempt to correct their foreign accent, as happened to me in one case. There is a view that learners are more likely to have accurate perceptions of phonological and lexical correction, and less likely have accurate perceptions of grammatical correction. This seems plausible given that the former usually only involve individual words or syllables, and so might be easier to process. Teachers can also supply both the 'correct' and the 'incorrect’ pronunciation to make the distinction clearer.

It is worth checking understanding during delayed correction, and it could be worth checking or explaining the nature of the error on-the-spot, if not too disruptive.

Why do learners repeat mistakes after they’ve been corrected?
It is difficult to change one’s L2 speaking and writing habits. It takes time for new language to be internalised. People might also feel shame around reading, recalling or listening back to their mistakes, making it harder to criticise and change them. That is why I emphasise that we should only be correcting learners to the extent that they show interest in being corrected.

Does correcting learner errors actually improve the accuracy of their language production?
Even if you do everything right, it is not inevitable. Learners have to pay attention to the corrections, perhaps understand them explicitly, and consciously work to incorporate them into speech and writing.

If it doesn’t, what other reasons could there be to correct errors?
Other reasons could be: it helps people build up their confidence in the L2 if they know that you will point out errors as they occur; it is useful for other learners who are listening; it shows learners that you're paying attention to their language use and taking an interest in their progress (although there are other ways to do this, such as by simply pointing out where they have made progress).

Criticism and response
One could argue that the idea of 'just do what students want' is a cop-out, for two reasons. Firstly, teachers should come with their own principles and views about what types of language knowledge and classroom activities students should value. This could include true ideas about the potential benefits of attending to corrections of one's mistakes, as well as the broader aspects of the language that they relate to. Secondly, teachers' views should be informed by research about the effects of feedback, not by what students say they want.

On the first point: teachers should try to persuade students of those views and principles in the course of their interaction, while taking into account individual differences. This, in itself, is an important principle. On the second point: the affective and motivational aspects discussed here override other considerations (such as the relative effectiveness of different types of feedback, or peer vs teacher feedback). Discussions of 'the effects of feedback' are meaningless unless they take into account the learner's own experience and views about the interaction. So, no, it is not a cop-out.
 

* See this handout describing interactional recasts and other types of feedback.

** More discussion of this and other examples in this paper (which is also the source of the extract above).

Mistakes and correction in language learning

Suppose you're trying to teach someone a new language. At some point they get good enough to have short exchanges with you. They're not a complete beginner anymore, but they still can't speak with the same ease, precision or complexity that you can. Almost inevitably, they make mistakes. These could be mistakes in grammar, choice of words, or pronunciation. They don't actively want to make mistakes, and in many cases they want you to point out their errors. So do you just stop them at every error and correct them?

Unless they insist on it, that is unlikely to be right. I do think one of the single best ways to help a person acquire correct grammar is simply to point out their errors, and explain them in terms that they can understand. But excessive interruption to correct errors may do more harm than good. While it might preserve some notion of grammatical purity in the classroom, it hinders other valuable processes that are going on in the learner's mind while he's involved in production of the target language. He's thinking about how to start the sentence, which verb forms to use, how to connect the different ideas together, how to pronounce the less familiar words, and so on. He may also be thinking about what he's trying to say - and it's hard to pay conscious attention to form if you haven't yet worked out the content.

So it's not obvious when to correct language learners on their errors. Here are some questions to start with:

• Should minor errors ever be corrected? (I mean things like wrong prepositions, word order in questions, missing articles, etc.) If so, why?
• How much context is needed for correction of an error to be meaningful or beneficial?
• Is it ever harmful to correct language learners' mistakes?
• How often do learners even know what you're correcting them on? Is it worth checking? And if so, do you need to check every time, or just in specific cases? 
• Why do learners repeat mistakes after they’ve been corrected on them?
• Does correcting learner errors actually improve the accuracy of their language production?
• If it doesn’t, what other reasons could there be to correct errors?

On the first point, a key argument is that errors only matter insofar as they affect communication — which would mean there’s no point in correcting what I’ve called ‘small’ errors. If the goal is to become an effective communicator, there are lots of things which are more important than grammatical accuracy per se (1). Effective communication depends on clear pronunciation, appropriate word choice, ability to connect ideas logically, knowledge of the culture’s social conventions, and more. Correcting the small stuff doesn’t necessarily help people become effective communicators. Even if it were important, there is a view (2) that grammatical correction doesn’t even lead to improvements in accuracy. This could be connected with general criticism of traditional approaches to teaching grammar — that is, as a core component of a language course that can be presented without much if any reference to meaning.

Still, I keep correcting students. So I’ve been asking myself the final question listed above. Why do I bother correcting them? I think it’s important to have a good answer to this question, because it is easy to fall into the mindset of: “I taught you this last week, so you ought to remember it now” — and then to despair when students repeat their mistakes. Answer (eventually) to follow.

Notes

1. Mistakes and Correction by Julian Edge is a good introduction to this topic.

2. E.g. Truscott 1996 - although I think this was only about correction of students' written work.

3. Another problem in this area is the following unpleasant paradox that I have observed in my own teaching practice and experienced as a learner:

 If an error is small, it's easy to correct it without causing confusion or taking up too much time. If an error is serious, it's harder to correct by the same criteria - especially if the teacher himself isn't sure what the learner is trying to say. But: If an error is small, it doesn't benefit the learner much to be told about it, for the reasons discussed in the post above. If an error is serious, it benefits the learner a lot to be told about it. So in practice, the more you stand to benefit from an error being brought to your attention, the less likely it is that this will happen.

(I am assuming that learners are developmentally ready for and interested to hear the corrections in question - otherwise they don't benefit either way.)