Recasting

A recast is a corrective move in which the teacher repeats a learner’s incorrect utterance in corrected form, without stopping the conversation or stating that an error was made. The meaning is kept, the form is quietly fixed, and the correction is implicit. It is the most frequently used correction technique in language classrooms.

Where the term comes from

The word entered language teaching through second language acquisition research rather than through teacher training. Michael Long’s work on the linguistic environment in the 1990s treated the recast as a form of negative evidence delivered inside meaningful talk, borrowing the idea from studies of how adults respond to small children, who are rarely told they are wrong and are frequently echoed back in corrected form.

The term became standard in classrooms through Roy Lyster and Leila Ranta’s 1997 study Corrective Feedback and Learner Uptake: Negotiation of Form in Communicative Classrooms, published in Studies in Second Language Acquisition 19(1), pages 37 to 66. They recorded 18.3 hours of interaction across 27 lessons in four primary French immersion classrooms in Montreal, and sorted every correction a teacher made into six categories. Those six categories are still the standard vocabulary for talking about correction, and the recast is the first of them.

Their headline finding was not that recasts are common. It was the combination: teachers showed an overwhelming tendency to use recasts in spite of the latter’s ineffectiveness at eliciting student-generated repair.

The numbers that made the term famous

All figures below are from Lyster and Ranta (1997) and describe their four immersion classrooms. They are the most quoted numbers in the correction literature and they are not a universal constant.

Feedback type	Share of all feedback moves	Leads to learner-generated repair
Recast	55%	None, by definition. The correct form has already been supplied, so there is nothing left for the learner to generate.
Elicitation	14%	45%
Clarification request	11%	27%
Metalinguistic feedback	8%	46%
Explicit correction	7%	None, by definition.
Repetition of error	5%	31%

Uptake, meaning any learner response at all following the feedback, ran at 31% after a recast, 50% after explicit correction and 100% after elicitation. More than half of all correction in those classrooms was delivered by the move that drew the least response.

How it is used in practice

A recast takes one turn and costs nothing in lesson time, which is a large part of why it dominates. It has three common shapes.

1. The full recast. The teacher repeats the whole utterance with the error fixed. Natural, invisible, and the easiest to miss.
2. The partial recast. The teacher repeats only the corrected fragment, which isolates the error and makes it far more noticeable. This is the version worth practising.
3. The stressed recast. The corrected element carries extra stress or a rising tone, which pushes an implicit move towards the explicit end without ever naming the mistake.

Translation counts as a recast too. When a learner reaches for their first language and the teacher supplies the English, the move is doing the same job: the meaning is accepted and the form is supplied.

Recasts are the default during fluency work, where stopping a speaker to analyse a verb ending costs more than the error does. They are a weaker choice during accuracy work, where the whole point of the stage is that the learner produces the form.

A worked classroom example

A B1 evening class of nine adults in Cork, thirty minutes into a speaking lesson about holidays. The teacher is not correcting much, deliberately, because the aim of the stage is fluency.

A learner says: I go to Spain last summer with my brother.

The teacher says: Ah, you went to Spain last summer.

The learner says: Yes, very hot. The conversation continues.

This is the characteristic failure of the technique, and it is not a failure of attention on the learner’s part. It is a design problem. Lyster’s 1998 follow-up study classified 377 recasts from the same corpus and compared them with the teachers’ non-corrective repetitions, where nothing was wrong and the teacher simply echoed a learner. The two moves turned out to serve identical functions in roughly equal proportions, and teachers delivered positive feedback such as good and yes alongside both of them at similar rates. From where the learner is sitting, a correction and an agreement can be acoustically and pragmatically the same event.

Two cheap repairs are available, and neither requires abandoning the technique.

Make the recast partial. Instead of returning the whole sentence, return the fragment: You went. Three syllables, the error isolated, and the flow of the conversation barely disturbed. The learner has something to compare against because only one thing came back.

Or switch to a prompt. You go, or you went. That is elicitation, it costs one extra turn, and it is the move the research says is most likely to get the learner to say the right form themselves.

Commonly confused with

Recast	Often confused with	The distinction
Recast	Explicit correction	Explicit correction states that the utterance was wrong and then supplies the form: Not go, we say went. A recast supplies the form and says nothing about the error. Both are input-providing, and neither leaves the learner anything to work out.
Recast	Prompts, meaning elicitation, metalinguistic feedback, clarification requests and repetition of error	This is the distinction that matters most. Prompts withhold the correct form and push the learner to produce it. Recasts hand it over. The whole argument about correction is really an argument about which of the two a given moment calls for.
Recast	Non-corrective repetition, or echoing	Echoing repeats a correct utterance to confirm or to let the rest of the class hear it. A recast repeats an incorrect one with the form changed. They sound the same from the back of the room, which is precisely the problem the research identifies.
Recast	Repetition of error	Repetition of error gives the mistake back unchanged, usually with a rising tone: You go? Nothing is supplied and the learner has to act. It is a prompt. A recast is its opposite.
Recast	Reformulation	In most writing, reformulation is the written cousin: a learner’s text rewritten by a proficient user so the two versions can be compared. It is read and studied rather than heard in passing, so it does not carry the ambiguity problem.
Recast	Eliciting	Eliciting draws language out of learners before anything has gone wrong, as a way of presenting or revising. Recasting responds to something that has already been produced incorrectly. The timing is the difference.

What the evidence actually supports

The recast is the one correction technique that is routinely described as both the most common and the least effective, and that summary is too simple in a way worth understanding.

Lyster and Ranta measured uptake and repair: what the learner did in the turn immediately after the feedback. On that measure the recast performs badly, and necessarily so, because a move that supplies the answer cannot also produce an answer. What it does not measure is acquisition. A learner who says nothing in the next turn may still have registered the difference.

Later meta-analytic work, including Lyster and Saito’s synthesis of oral feedback in classroom settings, found that corrective feedback as a whole does help, and that prompts tended to outproduce recasts in classroom studies while recasts held up better in laboratory conditions. Dan Brown’s meta-analysis of observational studies confirmed the frequency pattern across many more classrooms than the original four: reformulating moves, recasts chief among them, still account for roughly two thirds of what teachers do.

Three honest limits follow, and they are usually left off pages about this term.

• The famous numbers come from four classrooms of children in French immersion. Not adults, not English, not a language school. The six categories travel well. The percentages should be treated as a finding about those rooms rather than a law.
• Low uptake is not proof of low learning. It is proof of low visible response, which is what was counted. This page does not claim recasts fail to teach anything, because the study design cannot show that.
• Noticing is the live question. The evidence that recasts work is strongest where learners could tell a correction was happening. That is an argument for making recasts more salient, not for abandoning them.

The defensible claim is a narrow one. A recast is the right move when the priority is keeping a learner talking, and it becomes a wasted move when it is used so often, and so gently, that nobody in the room can tell it apart from agreement.

Sources and further reading

The ERIC record below was opened and checked on 9 October 2026. The remaining items are journal and print titles with no freely readable edition to link to, so they are cited bibliographically rather than linked.

Source	What it provides
Lyster, R. and Ranta, L. (1997), Corrective Feedback and Learner Uptake: Negotiation of Form in Communicative Classrooms, Studies in Second Language Acquisition 19(1), 37 to 66	The six-category taxonomy, the corpus of 18.3 hours across 27 lessons, and every percentage in the table above.
Lyster, R. (1998), Recasts, repetition, and ambiguity in L2 classroom discourse, Studies in Second Language Acquisition 20	The analysis of 377 recasts against non-corrective repetition, and the finding that the two serve identical functions in equal proportions.
Lyster, R. and Saito, K. (2010), Oral feedback in classroom SLA: a meta-analysis, Studies in Second Language Acquisition 32	The synthesis showing corrective feedback helps overall, and the prompt and recast comparison across classroom and laboratory settings.
Brown, D., The type and linguistic foci of oral corrective feedback in the L2 classroom: a meta-analysis, Language Teaching Research	Confirmation of the frequency pattern across far more classrooms than the original four, with reformulating moves at roughly two thirds of teacher feedback.
Long, M. H. (1996), The role of the linguistic environment in second language acquisition, in Ritchie, W. and Bhatia, T. (eds), Handbook of Second Language Acquisition, Academic Press	The account of recasts as negative evidence delivered inside meaningful interaction, which is where the classroom term came from.
Ur, P. (2012), A Course in Language Teaching, second edition, Cambridge University Press	The standard teacher-training treatment, which lists the recast alongside elicitation, clarification requests, metalinguistic feedback and explicit correction as the practical options for oral feedback.

Frequently asked questions

Is a recast the same as correcting a student?

It is one way of correcting, and the quietest one. A recast supplies the right form inside an ordinary-sounding reply and never says that anything was wrong. That is what separates it from explicit correction, which names the error first, and from prompts, which point at the error and leave the learner to fix it.

Do recasts actually work?

They work less visibly than the alternatives rather than not at all. The classic study measured what learners did in the very next turn, and on that measure recasts came last, because a move that hands over the answer cannot also produce one. Later meta-analyses found that corrective feedback does aid acquisition and that recasts perform better when learners can tell a correction is taking place.

When should I use a recast rather than a prompt?

Use a recast when the stage is about fluency and stopping the speaker would cost more than the error does, when the learner is fragile or new to the group, or when the item is beyond their current level and they could not produce it if asked. Use a prompt when the form is one the learner already knows and has simply slipped on, which is most of the time during accuracy work.

What is the difference between an implicit and an explicit recast?

It is a matter of degree rather than a hard line. A full recast delivered in a normal tone is at the implicit end and is easy to miss. Shortening it to the corrected fragment, adding stress to the changed element, or using a questioning tone pushes it towards the explicit end while still never stating that an error occurred.

Can recasts be used in written work?

The written equivalent is usually called reformulation: the learner’s text rewritten in correct English so the two can be compared side by side. It is more effective than a spoken recast for the simple reason that a written version sits still, and the learner can look from one to the other for as long as they need.

Why do teachers use recasts so much if they produce less repair?

Because they cost nothing. A recast takes one turn, breaks no one’s concentration, embarrasses nobody and lets the conversation carry on. Every other correction move costs time or face or both. The honest answer is that the recast is popular for the teacher’s reasons as much as the learner’s, which is why counting your own is a useful exercise.

Related terms in this glossary

• Interlanguage, the developing learner grammar that errors are evidence of, and the reason a recast is answering a system rather than a slip.
• L1 interference, the source of the errors most often recast, and the pattern that tells you whether one correction will be enough.
• Eliciting, the move that draws language out of learners, and the family that prompts belong to.
• The full TEFL teaching methods and classroom techniques glossary.

For the wider question of which errors to correct and when to leave them, see error correction in ESL on The TEFL Institute, and for what the qualification and career terminology behind classroom practice means, see TEFL qualifications and terminology explained. For how much correction practice a training course actually includes in Ireland and what employers look for, see 120 hour versus Level 5 TEFL on The TEFL Institute of Ireland.

Last updated 9 October 2026. Written by Katie Troy for tefl.ai. All linked external sources were opened and checked on that date.