← Garden Patch Home · Citations
Schegloff, Jefferson, and Sacks demonstrate that conversational repair — the practices participants use to address problems of speaking, hearing, and understanding — operates as a coherent system with structural preferences. The system is organized so that the speaker who produced the trouble source gets the earliest and most opportunities to correct it. Other-correction is systematically delayed and structurally dispreferred.
Repair taxonomy: Repair is classified along two dimensions — who initiates and who completes — yielding four types: self-initiated self-repair, self-initiated other-repair, other-initiated self-repair, and other-initiated other-repair. These are not equally distributed; the system favors self-initiation and self-completion.
Positional hierarchy: Self-repair opportunities are distributed across specific positions: within the trouble-source turn itself, in the transition space between turns, and in the third turn after an other-initiation. Each position has characteristic techniques (cut-offs, sound stretches, word replacements within the turn; “uh” and restarts in the transition space).
Other-initiation as invitation: When others initiate repair, they use techniques that invite the original speaker to self-correct — open class initiators (“huh?”), partial repeats, question-word repeats — rather than correcting directly. This preserves the structural preference even when others intervene.
Organizational not psychological preference: “Preference” here is a technical term from conversation analysis. The system provides structural priority to self-correction through turn-taking mechanics, not through individual motivation or politeness norms. The organization itself produces the preference.
Social consequences: The system protects speaker competence while maintaining mutual understanding. Direct other-correction, when it occurs, is typically modulated (marked with delay, agreement prefaces, or embedded in larger turns) precisely because the system treats it as dispreferred.
“The organization of repair in conversation provides for self-correction in a systematic way… The set of practices which make up the organization of repair are so organized as to favor self-correction.”
This paper established the analytical framework for all subsequent work on conversational repair, a subfield with thousands of studies across 50 years. The repair taxonomy and preference hierarchy remain the standard framework in conversation analysis, applied to L2 interaction, clinical settings, institutional talk, and human-computer interaction. For Deep Context Architecture, the repair organization provides a model for how agent protocols can handle errors: give the error-producing agent first opportunity to self-correct before escalating to other-correction.
Primary: Language, Vol. 53, No. 2 (1977), pp. 361-382
relates_to::[[Errors as Negotiation Opportunities]] The repair preference hierarchy demonstrates that errors in conversation are not failures but structured opportunities for participants to negotiate understanding. Self-correction preserves the error-producer’s agency; other-initiation creates a collaborative repair sequence.
relates_to::[[Conversational Repair Organization]] This paper is the primary source for the Conversational Repair Organization model. Schegloff, Jefferson, and Sacks define the taxonomy and preference organization that the model explores across different interaction contexts.