Error Correction in SLA

Definition:

Error correction in SLA — formally called corrective feedback (CF) — refers to any response to a learner’s non-target-like utterance that signals that an error has been made, with the purpose of promoting more accurate production over time. Corrective feedback operates in classrooms, conversation exchanges, online tutoring, and automated tools, and spans a spectrum from immediate explicit correction (“No, the correct form is…”) to subtle implicit signals (repetition of the learner’s utterance with the error fixed, without comment). Decades of research have examined whether error correction promotes acquisition at all (a long-contested question), which error types benefit most, which feedback strategies are most effective, and for whom. The current evidence broadly supports the value of focused, targeted corrective feedback over no feedback — but implementation details matter greatly for whether correction helps or merely disrupts communication.

Types of Corrective Feedback (Lyster & Ranta Taxonomy)

Roy Lyster and Lious Ranta’s 1997 taxonomy of CF types in classroom discourse remains foundational:

1. Explicit correction: Teacher directly provides the correct form with an indication that an error occurred (“No, you should say ‘I went,’ not ‘I goed’”)

1. Recast: Teacher reformulates the learner’s erroneous utterance correctly without explicitly marking it as an error (“Teacher: So you went there yesterday? [implicit correction of ‘goed’]”)

1. Clarification request: Teacher requests clarification that may signal a problem (“Sorry, what do you mean?”) — puts the burden of repair on the learner

1. Metalinguistic feedback: Teacher provides a grammatical hint or description without giving the correct form (“You need the past tense here”)

1. Elicitation: Teacher directly elicits the correct form from the learner (“How do we say that with the correct verb form?”)

1. Repetition: Teacher repeats the learner’s error with rising intonation to signal a problem

The Recasts Debate

Recasts — reformulating learner errors without explicit comment — are the most common form of CF in naturalistic interaction and classrooms. The debate:

• Pro-recasts: Recasts are conversationally natural, maintain communication flow, and may support implicit acquisition by providing target-like input immediately after an error
• Con-recasts: Research (Lyster, Saito) finds that learners frequently fail to notice recasts as corrective — they interpret reformulations as conversational confirmation rather than correction. This limits their efficacy compared to more explicit feedback

Explicit vs. Implicit Corrective Feedback

Meta-analyses (Li, 2010; Norris & Ortega, 2000; Russell & Spada, 2006) consistently find:

• Both explicit and implicit CF outperform no feedback for the treated features
• Explicit CF tends to produce stronger immediate effects — learners are more likely to notice and uptake explicit correction
• Implicit CF may support more durable acquisition in some conditions — but the evidence is mixed
• Focused CF (targeting one or two error types systematically) outperforms unfocused CF (correcting all errors unpredictably)

Which Errors to Correct

Not all error types benefit equally from correction:

• Treatable errors (systematic rule-based errors — third-person -s, article use, past tense) respond well to corrective feedback
• Non-treatable errors (spelling, pronunciation, vocabulary choice) respond less systematically to grammatical correction
• Salient, focused correction of one specific error type across a period of time produces more measurable acquisition than random correction of all errors

The Affective Side

Research consistently shows that fear of being corrected increases speaking anxiety — learners who feel corrections are threatening tend to over-monitor and produce less output. Effective CF delivery is calm, matter-of-fact, and targeted — not pointing to every error as a failure.

History

1985 — Krashen’s anti-correction argument. Krashen argues that error correction has minimal effect on acquisition; the Monitor can only apply explicit knowledge and only in conditions of slow, careful production.

1997 — Lyster and Ranta. Taxonomy of CF types and analysis of classroom corrective feedback uptake; finds recasts produce low uptake rates despite being most common.

2000 — Norris and Ortega meta-analysis. Explicit instruction (including error correction) produces measureable effects on L2 acquisition; challenges the Krashenian anti-instruction position.

2004 — Long, Interaction Hypothesis. Argues that negotiated interaction including corrective feedback is a key mechanism for L2 acquisition.

2010 — Li meta-analysis. Compares CF types across studies; confirms explicit > implicit for immediate outcomes; mixed for delayed outcomes.

Common Misconceptions

“Correcting every error improves learning.” Research consistently shows that over-correction — providing corrective feedback on every error — can impede acquisition by overwhelming the learner’s capacity to process and attend to feedback, creating anxiety that inhibits output, and preventing learners from noticing patterns in their own interlanguage. Selective, focused corrective feedback on pedagogically targeted forms outperforms comprehensive error correction in acquisition outcomes.

“Learners know they’ve been corrected when given recasts.” Recasts (reformulations of learner utterances in target-like form) are the most common corrective feedback type in naturalistic conversations, but research has shown that learners frequently perceive recasts as conversational confirmation rather than correction — particularly in meaning-focused, communicative interactions where corrective intent is ambiguous. This “shielded” quality of recasts means they are often not processed as correction.

Criticisms

Error correction research has produced mixed and sometimes contradictory findings, leading to significant uncertainty about optimal feedback practices. The ecological validity of experimental error correction studies has been questioned — laboratory settings controlling feedback type do not capture the complexity of real classroom interactions where multiple feedback types co-occur. The debate between explicit and implicit form-focused instruction (of which error correction is one dimension) remains unresolved, with meta-analyses showing advantages for explicit feedback in controlled studies but limitations in generalizing to naturalistic L2 development.

Social Media Sentiment

Error correction is one of the most contested topics in language teaching methodology communities online. Teachers and learners debate whether and when to correct spoken errors, with strong opinions on both sides — some teachers advocate for immediate explicit correction; others follow communication-first principles and correct only after communication goals are met. Language learning communities have nuanced discussions about self-correction, peer-correction, and tutor correction that reflect awareness of the research findings without always connecting them to the academic literature. The frustration of receiving unhelpful correction (e.g., constant interruptions during speaking practice) is a frequently shared learner experience.

Last updated: 2026-04

Practical Application

1. Ask tutors and language partners to correct specific error types. “Please correct every time I use the wrong verb tense” — focused, systematic correction is far more effective than random error flagging.

1. Welcome recasts but learn to recognize them. When a conversation partner repeats your sentence with a correction, consciously register it as feedback — don’t just hear it as agreement.

1. Don’t expect conversation partners to fix everything. Most people are polite; they won’t correct every error. Structured feedback (tutors, written correction, apps) is more reliable for systematic improvement.

Related Terms

• Focus on Forms
• Explicit Learning
• Noticing Hypothesis
• Interaction Hypothesis
• Speaking Anxiety

See Also

• Focus on Forms — Pedagogy that incorporates corrective feedback within communicative tasks
• Noticing Hypothesis — Cognitive basis for why correction that prompts noticing drives acquisition
• Speaking Anxiety — Affective dimension affected by how correction is delivered
• Sakubo

Research

Long, M. H. (1996). The role of the linguistic environment in second language acquisition. In W. C. Ritchie & T. K. Bhatia (Eds.), Handbook of Second Language Acquisition (pp. 413-468). Academic Press.

Introduces the Interaction Hypothesis and formalizes the distinction between implicit and explicit negative feedback (recasts vs. explicit correction), providing the theoretical basis for understanding error correction as one mechanism among several through which interaction supports acquisition.

Lyster, R., & Ranta, L. (1997). Corrective feedback and learner uptake. Studies in Second Language Acquisition, 19(1), 37-66.

A highly influential classroom study documenting the frequency and uptake of different error correction types in French immersion classrooms, demonstrating that recasts — despite being most frequent — produced less uptake than negotiation-based feedback types, reshaping subsequent research on optimal error correction strategies.

Mackey, A., & Goo, J. (2007). Interaction research in SLA: A meta-analysis and research synthesis. In A. Mackey (Ed.), Conversational Interaction in Second Language Acquisition (pp. 407-452). Oxford University Press.

A meta-analysis of interaction studies including corrective feedback research, synthesizing effect sizes across studies and documenting the conditions under which different corrective feedback types are most effective — essential for calibrating the practical implications of error correction research.