Repair Target Selection¶
Decision procedure — instantiates Locution-Illocution-Perlocution Decomposition
Given a diagnosed cross-layer mismatch, decides which single layer actually broke — wording, standing, or effect — and picks the matched remedy at the right intensity, so effort lands on the failing layer rather than the loudest one.
Once an utterance has been split into layers and their mismatch diagnosed, someone still has to decide where to intervene. Repair Target Selection is that decision procedure. It takes the layered diagnosis and answers two questions: which single layer is the one that actually broke, and what remedy — at what intensity — is matched to it. Its defining move is matching: a wording problem gets a wording fix, a standing problem gets a role-or-channel fix, an effect problem gets mitigation, and the mechanism's whole value is refusing to fix the wrong layer just because it is the most visible. Because so many communication breakdowns are really authority breakdowns — the act didn't count because the speaker lacked standing, used the wrong channel, or violated a convention — it builds the roles-and-standing picture explicitly, as the pivotal candidate repair site. It selects and prescribes; it does not parse words, classify force, or trace effects.
Example¶
An engineering team runs an autonomous coding agent. Told "clean up the old logs directory," the agent deleted a directory that a live production pipeline was actively writing to, and a service went down. The reflexive fix is to reword the instruction. Repair Target Selection resists that until it has localized the break. It reviews the layered diagnosis. At the wording layer, "old logs directory" was mildly ambiguous — but ambiguity alone did not cause the outage. At the standing/authority layer sits the real failure: the agent treated a cleanup request as an authorization to delete production infrastructure, and no convention required it to confirm before acting on a production path. It had the capability but not the warranted standing, and nothing in the setting supplied the missing check. At the effect layer, the outage itself needs mitigation.
The selection follows from that localization. The primary repair targets standing: scope the agent's delete authority, and require explicit confirmation for production paths — a procedural correction, not a reworded prompt. A secondary repair addresses the effect: restore from backup, add a guardrail, run a postmortem. The wording tweak is real but minor. What the mechanism prevents is the tempting non-fix — rewording the instruction while leaving the authorization hole wide open for the next ambiguous request.
How it works¶
- Take the layered diagnosis as input. The cross-layer mismatch — intended versus recognized versus actual — arrives from Effect-Chain Tracing and the single-layer analyzers; this mechanism consumes it rather than re-deriving it.
- Localize to the failing layer. Identify which layer's breakdown is causal, not merely present, and distinguish the primary failure from downstream symptoms.
- Build the standing picture for the authority candidate. Reconstruct the roles, norms, channel, and conventions to judge whether the act failed because it did not count — the layer most often misattributed.
- Match remedy to layer, and set intensity. Wording repair, force restatement, role or standing clarification, effect mitigation, procedural correction, or accountability action — chosen to fit the diagnosed layer and scaled to the stakes.
- Sequence, don't over-repair. Order primary and secondary repairs without patching layers that did not break.
The distinguishing discipline is that it is a decision, downstream of diagnosis and upstream of execution: it names the target and the remedy, and stops there.
Tuning parameters¶
- Localization resolution — whether you commit to one failing layer or allow a small ranked set. Single-layer focus keeps the repair sharp but can miss a genuine two-layer break; multi-layer selection is thorough but risks diffuse, expensive fixes.
- Repair intensity — how strong a remedy the layer warrants, from a one-line clarification up through restatement, procedural correction, and formal accountability action. Under-powered repairs leave the break open; over-powered ones burn trust and time on a small fault.
- Authority sensitivity — how strongly standing, channel, and convention weigh as the candidate failing layer. Set too low, real authorization failures get mis-repaired as wording; set too high, ordinary phrasing slips get escalated into standing disputes.
- Over-repair guard — how strictly you refuse to fix layers that did not break. Tight guarding keeps effort targeted; loose guarding fixes everything and learns nothing about what actually failed.
When it helps, and when it misleads¶
Its strength is that it defeats the archetype's most wasteful reflex — repairing the loudest layer instead of the broken one. Rewording a message that failed on authority, or apologizing for an effect that a standing fix would prevent recurring, both feel like action while leaving the real fault in place; matched repair is what converts a diagnosis into a fix that holds.
It misleads when the localization is wrong. Mis-localize, and you aim a precise remedy at the wrong layer — confidently. The archetype's authority-condition blindness is the classic version: ignoring standing and channel, so an authorization failure gets papered over as a phrasing problem and recurs. The self-serving misuse is choosing the cheapest layer to fix rather than the one that broke — a quick reword to close a ticket that a harder standing change would actually resolve. The discipline that keeps it honest is to match the repair to the diagnosed layer rather than the convenient one, to treat authority as a live candidate on every pass, and to scale intensity to stakes rather than to appetite.[1]
How it implements the components¶
Repair Target Selection realizes the archetype's decision-and-remedy components — the ones that turn a layered diagnosis into an aimed fix:
authority_and_convention_context— reconstructs the roles, norms, channel, and conventions that determine whether the act counted, so a standing failure can be recognized as the repair site rather than misread as a wording slip.layer_specific_repair_plan— outputs the remedy matched to the diagnosed layer and scaled in intensity: reword, restate, clarify standing, mitigate effect, or escalate to procedural correction.
It does not trace what the utterance caused or compare intended against actual effect (perlocutionary_effect_trace, uptake_alignment_record — Effect-Chain Tracing), and it does not preserve or split the utterance into said- and done-layers (utterance_capture_record, locutionary_content_map, illocutionary_force_map — Layer Separation Reframing); it decides the fix, not the diagnosis. Deep testing of whether the standing conditions actually hold is context-condition testing, a same-named neighbor mechanism.
Related¶
- Instantiates: Locution-Illocution-Perlocution Decomposition — Repair Target Selection is the step that turns the layered diagnosis into an aimed, layer-matched remedy.
- Consumes: Effect-Chain Tracing supplies the intended-versus-actual mismatch this selection localizes and repairs against.
- Sibling mechanisms: Layer Separation Reframing · Effect-Chain Tracing · Literal Content Parsing · Force Classification · Context-Condition Testing · Uptake Confirmation
Editorial Notes¶
Form Classification¶
Form family: Decision, Gate & Allocation
Rationale: Repair Target Selection operates as a case-specific gate, selection, routing, prioritization, or resource disposition because it given a diagnosed cross-layer mismatch, decides which single layer actually broke — wording, standing, or effect — and picks the matched remedy at the right intensity, so effort lands on the failing layer rather than the loudest one.
Independent corroboration: The frozen evidence defines Repair Target Selection as 'Given a diagnosed cross-layer mismatch, decides which single layer actually broke — wording, standing, or effect — and picks the matched remedy at the right intensity, so effort lands on the failing layer rather than the loudest one', so its operative form is Decision, Gate & Allocation.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Linguistics & Semiotics
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Universal
Rationale: Layer-specific repair descends from conversation analysis and speech-act distinctions among wording, force, and uptake.
Related originating lineages:
- Philosophy — Philosophy of language supplied speech-act conditions of authority, standing, and successful performance.
- Rhetoric — Rhetorical practice materially shaped audience- and effect-matched correction.
Review resolution: Both blind reviewers agree that linguistics_semiotics is the primary historical origin. Explicit reconciliation of domain reach disagreement adopts reviewer_a's evidence: Layer-specific repair descends from conversation analysis and speech-act distinctions among wording, force, and uptake. The selected record uses alternates=philosophy, rhetoric, origin_mode=cross_disciplinary_synthesis, and domain_reach=universal; the other review proposed alternates=philosophy, rhetoric, origin_mode=cross_disciplinary_synthesis, and domain_reach=multi_domain. The selected combination better preserves the mechanism-specific formative lineages and calibrated scope; broader present-day use is not treated as proof of additional historical origin.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
This is the archetype's reconnection point. The layers are separated for analysis, but the repair decision is where they rejoin: choosing the target means holding words, act, standing, and effect side by side and judging which one, fixed, resolves the whole. Keeping selection distinct from execution lets the same diagnosis support anything from a quiet reword to a formal accountability action without re-opening it.
References¶
[1] Schegloff, Emanuel A., Gail Jefferson, and Harvey Sacks. "The Preference for Self-Correction in the Organization of Repair in Conversation". Language 53(2): 361-382, 1977. Shows that other-initiated conversational repair first locates a specific trouble source and then gives the original speaker an opportunity to repair it; it does not address authority checks or scaling intensity to stakes. registry ↩