Browse the wiki

Output and corrective feedback

From the YapLearn Wiki · Learning method · Updated 2026-10-28

Output is the language a learner produces, in speech or writing. Corrective feedback is any response that tells a learner, directly or indirectly, that something said or written was not correct in the new language [1]. Two lines of research hold that both matter for learning: Merrill Swain's output hypothesis argues that producing language makes learners notice gaps in what they know, and Michael Long's interaction hypothesis argues that conversation, and especially the work of resolving misunderstandings, helps acquisition [2][3].

Corrective feedback has been measured in classrooms and laboratories since the 1990s. Several meta-analyses (studies that pool the results of many earlier studies) found that feedback helps learners develop more accurate language and that the effect lasts on delayed tests [4][5][6]. Feedback that pushes the learner to correct the error, rather than simply supplying the right form, tends to have the larger effect in classrooms [6].

Contents

Key facts

  • Swain developed the output hypothesis from research on Canadian French immersion students, who understood French well but spoke and wrote it with persistent errors [2].
  • Long's 1996 interaction hypothesis proposed that negotiating meaning in conversation connects input, the learner's attention and output in ways that support acquisition [3].
  • Lyster and Ranta's 1997 study of four primary immersion classrooms, based on 18.3 hours of recorded lessons, identified six types of oral feedback; teachers used recasts far more than any other type, although recasts rarely led students to correct themselves [1].
  • Li's 2010 meta-analysis of 33 studies found a medium overall effect of corrective feedback that was maintained over time; implicit feedback held up better on delayed tests than explicit feedback [5].
  • Lyster and Saito's 2010 meta-analysis of 15 classroom studies with 827 learners found that feedback had significant, durable effects, larger for prompts than for recasts and larger for younger learners [6].
  • Bloom reported in 1984 that students taught one to one with corrective feedback and mastery checks performed about two standard deviations above students in conventional classes [7].

Output and interaction

Output hypothesis

Origins in French immersion

The output hypothesis grew out of research on Canadian French immersion programs (see Immersion and study abroad), in which English-speaking children are taught school subjects in French. After years of input, immersion students understood French well, yet their speech and writing still showed many grammatical errors [2]. Swain argued that understanding a message does not require attention to every grammatical detail, while producing one does, and that immersion students had too few chances to speak at length and were rarely pushed to be accurate [2].

Functions of output

Swain proposed that output does several things that input alone does not. Producing language makes learners notice a gap between what they want to say and what they can say; it lets them test hypotheses about how the language works; and talking about their own language helps them reflect on it [2][8]. In a study with Sharon Lapkin, immersion students thought aloud while writing in French; when they ran into a problem, they often stopped to work out the form, a process the authors argued could generate new knowledge or strengthen existing knowledge [8].

Output compared with repetition

Output in this sense means producing original language. Repeating a model aloud, as in Shadowing, practices pronunciation and fluency but does not require the learner to build sentences. The output hypothesis is often contrasted with Stephen Krashen's view that Comprehensible input is the main cause of acquisition. Swain's own position was that input is necessary but not sufficient for accurate production [2].

Interaction hypothesis

Negotiation of meaning

Long's interaction hypothesis focuses on conversation between a learner and a more proficient speaker. When communication breaks down, speakers negotiate meaning: they ask for clarification, check understanding and rephrase. Long argued that this negotiation makes input more understandable and draws the learner's attention to forms that matter, and that feedback received during such exchanges, including recasts, can help learners notice differences between their own speech and the target language [3].

Input and output combined

In this view, interaction combines the benefits of input and output. The learner hears language adjusted to their level, produces language and gets immediate information about whether it worked [3].

Corrective feedback

Types of corrective feedback

Lyster and Ranta's classification, based on French immersion classrooms, remains the standard list of oral feedback types [1]. Their main finding was that teachers relied overwhelmingly on recasts, even though recasts were the type least likely to lead to student-generated repair, meaning a correction made by the student rather than the teacher [1]. Four types (elicitation, metalinguistic feedback, clarification requests and repetition) led to student-generated repair more often; the authors called the exchange these types start the negotiation of form [1]. The researchers measured uptake, a student's immediate response to feedback, and repair, an uptake that fixes the error.

The examples follow Lyster and Ranta's six types and respond to the learner error *Yo soy cansado*; they use the Ser and estar distinction, a common source of error for English speakers. Elicitation, metalinguistic feedback, clarification requests and repetition are grouped as prompts, because they push the learner to correct the error; recasts and explicit correction supply the correct form [6].

Six types of oral corrective feedback
TypeDescriptionExample teacher response
RecastTeacher repeats the sentence with the error corrected, without saying it was wrong*Ah, estás cansado.*
Explicit correctionTeacher states that it was wrong and gives the correct form*No se dice "soy cansado", se dice "estoy cansado".*
ElicitationTeacher prompts the learner to produce the form, for example by pausing or asking*Yo... ¿?*
Metalinguistic feedbackTeacher comments on the form without giving it*Para estados usamos otro verbo.*
Clarification requestTeacher signals the message was not understood or was wrong*¿Perdón?*
RepetitionTeacher repeats the error, usually with rising intonation*¿Soy cansado?*

Evidence from pooled studies

Russell and Spada 2006

Russell and Spada's 2006 meta-analysis pooled studies of corrective feedback on grammar, in speech and writing, and found a large overall benefit of feedback compared with no feedback [4].

Li 2010

Li's 2010 meta-analysis used different inclusion rules and covered 33 studies, 22 published and 11 doctoral dissertations [5]. It found:

  • A medium overall effect, maintained over time [5].
  • Better maintenance for implicit feedback, such as recasts, than for explicit feedback on delayed tests [5].
  • Larger effects in laboratory studies than in classrooms, in shorter treatments than in longer ones, and in foreign language settings than in second language settings [5].

Lyster and Saito 2010

Lyster and Saito's 2010 meta-analysis looked only at 15 classroom studies with 827 learners. Feedback had significant and durable effects; prompts had larger effects than recasts; effects were clearest on tasks in which learners produced free responses; and younger learners benefited more than older ones [6].

One-to-one correction

Bloom's tutoring results

Individual feedback is one reason one-to-one tutoring has long been considered unusually effective. In his 1984 "2 Sigma" paper, Benjamin Bloom reported that students taught by a tutor, with regular checks and corrective feedback until they mastered each unit, performed about two standard deviations above students in conventional group classes, so that the average tutored student outperformed about 98% of the comparison group [7]. Bloom's results came from general school subjects, not language learning; the approach is described under Mastery learning.

Feedback tailored to each learner

A tutor can give feedback tailored to each learner's errors, choose between recasts and prompts and track which errors recur, which a teacher with a full class rarely has time to do. Software and AI tutors aim to provide some of this individual feedback in conversation practice. The feedback research above was carried out with human teachers and researchers, and the reviews did not test automated feedback.

See also

References

  1. Lyster, R., Ranta, L. (1997). Corrective Feedback and Learner Uptake: Negotiation of Form in Communicative Classrooms. Studies in Second Language Acquisition, 19(1), 37-66. doi:10.1017/S0272263197001034
  2. Swain, M. (1993). The Output Hypothesis: Just Speaking and Writing Aren't Enough. The Canadian Modern Language Review, 50(1), 158-164. doi:10.3138/cmlr.50.1.158
  3. Long, M. H. (1996). The Role of the Linguistic Environment in Second Language Acquisition. Handbook of Second Language Acquisition (Academic Press), 413-468. doi:10.1016/B978-012589042-7/50015-3
  4. Russell, J., Spada, N. (2006). The effectiveness of corrective feedback for the acquisition of L2 grammar: A meta-analysis of the research. Synthesizing Research on Language Learning and Teaching (John Benjamins), 133-164. doi:10.1075/lllt.13.09val
  5. Li, S. (2010). The Effectiveness of Corrective Feedback in SLA: A Meta-Analysis. Language Learning, 60(2), 309-365. doi:10.1111/j.1467-9922.2010.00561.x
  6. Lyster, R., Saito, K. (2010). Oral Feedback in Classroom SLA. Studies in Second Language Acquisition, 32(2), 265-302. doi:10.1017/S0272263109990520
  7. Bloom, B. S. (1984). The 2 Sigma Problem: The Search for Methods of Group Instruction as Effective as One-to-One Tutoring. Educational Researcher, 13(6), 4-16. doi:10.3102/0013189X013006004
  8. Swain, M., Lapkin, S. (1995). Problems in output and the cognitive processes they generate: A step towards second language learning. Applied Linguistics, 16(3), 371-391. doi:10.1093/applin/16.3.371

Tags: Speaking · Feedback and correction · AI and one-to-one tutoring · Grammar

Output and corrective feedback | YapLearn Wiki