Repair Strategies

Definition:

Repair refers to the conversational practices by which speakers address problems in speaking, hearing, or understanding as they arise in interaction. Introduced as a technical term in Conversation Analysis by Schegloff, Jefferson, & Sacks (1977), repair is a fundamental mechanism of spoken interaction: conversations are perpetually susceptible to breakdowns (mishearing, misunderstanding, error, trouble in production), and repair is the collaborative system by which participants restore intersubjectivity—shared understanding—when it breaks down. In SLA, repair is studied as a window into interlanguage development, corrective feedback mechanisms, and the acquisition of interactional competence.

In-Depth Explanation

Conversation Analysis framework of repair:

Schegloff, Jefferson, & Sacks (1977) classified repair along two dimensions:

1. Initiation (who identifies the trouble):

Self-initiation: The speaker of the trouble-source initiates repair.
Other-initiation: A recipient initiates repair (indicating a problem).

2. Completion (who resolves the trouble):

Self-repair: The trouble-source speaker provides the correction.
Other-repair: The recipient provides the correction.

These two dimensions combine to produce four repair types:

Self-initiated self-repair (SI-SR): Speaker notices own error and corrects it (“I went to the — I went to the store”).
Self-initiated other-repair (SI-OR): Speaker requests help (“What’s the word for…?”).
Other-initiated self-repair (OI-SR): Recipient signals trouble; speaker repairs (“Are you coming? / — Oh, I didn’t mean Tom’s party, I meant…”).
Other-initiated other-repair (OI-OR): Recipient both signals and corrects the trouble — this is the least common type in ordinary conversation.

Preference for self-repair:

A consistently documented finding across languages is that self-repair is preferred over other-repair—speakers minimize direct correction of others and provide trouble-holders the opportunity to self-repair first. Other-repair is delayed, softened, or replaced by other-initiation sequences (giving the trouble-holder a chance to self-correct). This structural preference reflects broader face-work norms: directly correcting someone is an FTA (Face-Threatening Act).

SLA applications of repair:

Recasts as covert other-repair:

Lyster & Ranta (1997) classify different types of corrective feedback in L2 classrooms, including recasts—teacher reformulations of a learner’s erroneous utterance that provide the correct form without explicitly marking it as an error:

Learner: I goed to the store.
Teacher: You went to the store? What did you buy?

Recasts are a form of other-initiated other-repair, but they are embedded in the ongoing conversation (making them difficult to perceive as correction). Whether learners uptake recasts as corrections or simply as conversational continuations is a major empirical question (Lyster, 1998).

Pushed output through other-initiation:

Clarification requests and comprehension checks in L2 interaction create other-initiation sequences that push learners toward self-repair—the learner must identify and resolve the trouble in their own production. These are Swain’s (1985) pushed output opportunities via interactional pressure.

L2 learner repair:

L2 learners’ repair behavior differs from native speakers:

Higher rate of self-repair due to production difficulty (searching for words, monitoring output).
Communication strategy use as an alternative to repair: paraphrase, circumlocution, appeal for help.
Greater reliance on other-initiation for vocabulary trouble (“How do you say…?” / “What does X mean?”).
Interruption of repair sequences by face concerns (reluctance to signal non-understanding in high power-differential interactions).

Repair in Japanese conversation:

Japanese conversation repair exhibits several culturally-specific features:

Aizuchi (相づち): The dense backchanneling of aizuchi signals comprehension and prevents other-initiation repair sequences from being needed (or prolongs the signal of understanding). Absent aizuchi can signal comprehension trouble and trigger other-initiation.
Repair and indirection: Japanese indirect communication norms make direct other-initiation repair (explicitly saying “I don’t understand”) face-threatening; repair initiation is often oblique or delayed.
Keigo repair: Register-inappropriate forms may be repaired by native speakers via reformulation in the appropriate register — a form of keigo correction that is also other-initiated repair.

Communication strategies and repair:

When repair is not available (the learner can’t find the word), communication strategies are employed (paraphrase, mime, L1 insertion). These are not technically repair but alternatives to initiating a repair sequence—and their use is a central topic in SLA communicative competence research.

History

1977: Schegloff, Jefferson, & Sacks foundational CA repair paper.
1984: Pica & Doughty apply interaction/negotiation framework to L2 context.
1997: Lyster & Ranta classify L2 classroom corrective feedback types including recasts and repair-like moves.
1998: Lyster’s uptake study; challenges recast effectiveness as repair signal.
2001: Mackey, Gass, & McDonough experimental studies of interactional feedback and repair.
2004: Seedhouse’s The Interactional Architecture of the Language Classroom analyzes classroom repair systematically.
2010s: Conversation Analysis–SLA interface papers proliferate; repair studied in naturalistic L2 interaction.

Common Misconceptions

“Teachers should correct all learner errors.” CA research shows direct other-repair is the least preferred repair structure in conversation; immediate teacher correction may impede naturalness and doesn’t align with how repair functions in authentic interaction.

“Recasts are clear corrections that learners will always uptake.” Recasts are structurally ambiguous between correction and conversational continuation; research shows L2 learners frequently miss the corrective intent of recasts, especially in noise or when focused on content.

“Self-repair means the learner is struggling.” Self-repair is normative in all fluent speech; it signals monitoring and is not a sign of incompetence. High self-repair rates in L2 are evidence of active self-monitoring, which is positively related to accuracy development.

Criticisms

Most SLA repair research uses classroom interaction data, which has special institutional repair norms (IRE sequences, teacher authority) that diverge from naturalistic conversation.
The CA emphasis on sequential structure undertheorizes what repair reveals about underlying L2 representations vs. just conversational management.
Operationalizing “uptake” of repair (whether correction leads to acquisition) remains methodologically challenging.

Social Media Sentiment

Repair strategies are practically experienced by language learners in every conversation. “I used the wrong word and my Japanese partner just… repeated it differently” — learners describe confusion about whether a reformulation was a correction or just a natural continuation. The AJATT community’s emphasis on “shadowing rather than speaking with native speakers” partly reflects discomfort with the repair sequences and face management challenges of L2 conversation. Learners in Japanese immersion programs describe gradually learning the Japanese repair sequence conventions (how to signal non-understanding politely, how to self-repair without losing face).

Last updated: 2026-04

Practical Application

Embrace other-initiation as comprehension practice: When you hear something you don’t understand in Japanese, practice the formulaic other-initiation responses: もう一度言っていただけますか (could you say that again?), すみません、ちょっと聞こえなかったのですが (sorry, I didn’t quite hear that). Fluent other-initiation repair signals interactional competence even when comprehension fails.
Self-repair in output: Say what you mean, notice when it comes out wrong, correct overtly (Japanese conversation partners recognize this as normal). Don’t avoid complex forms for fear of repairing — repair is evidence of monitoring, which supports acquisition.
Recast monitoring: In tutoring or conversation exchange, specifically listen for whether a native speaker reformulates something you said — this is often a subtle recast. Attend to it; use it as a noticing-the-gap moment.
Aizuchi in Japanese conversations: Practice aizuchi (うん、なるほど、へえ、そうですか) to signal comprehension and reduce pressure for other-initiated repair sequences that would otherwise be triggered by your silence.

Related Terms

Research

Schegloff, E. A., Jefferson, G., & Sacks, H. (1977). The preference for self-correction in the organization of repair in conversation. Language, 53(2), 361–382. [Summary: Foundational CA repair paper; establishes preference for self-repair; describes four repair types and sequential organization; basis for all subsequent interactional and SLA repair research.]

Lyster, R., & Ranta, L. (1997). Corrective feedback and learner uptake: Negotiation of form in communicative classrooms. Studies in Second Language Acquisition, 19(01), 37–66. [Summary: Classifies six types of corrective feedback in L2 classrooms including recasts; documents uptake rates; foundational for corrective feedback/repair research in instructed SLA.]

Pica, T. (1994). Research on negotiation: What does it reveal about second-language learning, conditions, processes, and outcomes? Language Learning, 44(3), 493–527. [Summary: Comprehensive review of negotiation of meaning research; connects interactional repair sequences to L2 input modification and acquisition; foundational synthesis.]

Seedhouse, P. (2004). The Interactional Architecture of the Language Classroom: A Conversation Analysis Perspective. Blackwell. [Summary: CA analysis of classroom repair organization; distinguishes institutional repair norms from ordinary conversation; addresses pedagogical implications of repair structures in L2 classrooms.]

Mackey, A. (1999). Input, interaction, and second language development: An empirical study of question formation in ESL. Studies in Second Language Acquisition, 21(04), 557–587. [Summary: Experimental study of interactional feedback; demonstrates that other-initiated repair sequences producing pushed output are associated with subsequent question formation development; links repair to acquisition evidence.]