Governance artifacts
Governance files brought into scope by this page
This page is anchored to published surfaces that declare identity, precedence, limits, and the corpus reading conditions. Their order below gives the recommended reading sequence.
Q-Ledger JSON
/.well-known/q-ledger.json
Machine-first journal of observations, baselines, and versioned gaps.
- Governs
- The description of gaps, drifts, snapshots, and comparisons.
- Bounds
- Confusion between observed signal, fidelity proof, and actual steering.
Does not guarantee: An observation surface documents an effect; it does not, on its own, guarantee representation.
Q-Metrics JSON
/.well-known/q-metrics.json
Descriptive metrics surface for observing gaps, snapshots, and comparisons.
- Governs
- The description of gaps, drifts, snapshots, and comparisons.
- Bounds
- Confusion between observed signal, fidelity proof, and actual steering.
Does not guarantee: An observation surface documents an effect; it does not, on its own, guarantee representation.
Definitions canon
/canon.md
Canonical surface that fixes identity, roles, negations, and divergence rules.
- Governs
- Public identity, roles, and attributes that must not drift.
- Bounds
- Extrapolations, entity collisions, and abusive requalification.
Does not guarantee: A canonical surface reduces ambiguity; it does not guarantee faithful restitution on its own.
Complementary artifacts (1)
These surfaces extend the main block. They add context, discovery, routing, or observation depending on the topic.
Public AI manifest
/ai-manifest.json
Structured inventory of the surfaces, registries, and modules that extend the canonical entrypoint.
Evidence layer
Probative surfaces brought into scope by this page
This page does more than point to governance files. It is also anchored to surfaces that make observation, traceability, fidelity, and audit more reconstructible. Their order below makes the minimal evidence chain explicit.
- 01Canon and scopeDefinitions canon
- 02Weak observationQ-Ledger
- 03Derived measurementQ-Metrics
- 04Audit reportIIP report schema
Definitions canon
/canon.md
Opposable base for identity, scope, roles, and negations that must survive synthesis.
- Makes provable
- The reference corpus against which fidelity can be evaluated.
- Does not prove
- Neither that a system already consults it nor that an observed response stays faithful to it.
- Use when
- Before any observation, test, audit, or correction.
Q-Ledger
/.well-known/q-ledger.json
Public ledger of inferred sessions that makes some observed consultations and sequences visible.
- Makes provable
- That a behavior was observed as weak, dated, contextualized trace evidence.
- Does not prove
- Neither actor identity, system obedience, nor strong proof of activation.
- Use when
- When it is necessary to distinguish descriptive observation from strong attestation.
Q-Metrics
/.well-known/q-metrics.json
Derived layer that makes some variations more comparable from one snapshot to another.
- Makes provable
- That an observed signal can be compared, versioned, and challenged as a descriptive indicator.
- Does not prove
- Neither the truth of a representation, the fidelity of an output, nor real steering on its own.
- Use when
- To compare windows, prioritize an audit, and document a before/after.
IIP report schema
/iip-report.schema.json
Public interface for an interpretation integrity report: scope, metrics, and drift taxonomy.
- Makes provable
- The minimal shape of a reconstructible and comparable audit report.
- Does not prove
- Neither private weights, internal heuristics, nor the success of a concrete audit.
- Use when
- When a page discusses audit, probative deliverables, or opposable reports.
Interpretive persistence audit after deletion, correction, or 404
This framework addresses a case that has become central in generative environments: a source loses availability or primacy, yet its effects continue to appear inside outputs.
The classic mistake is to jump too quickly to “model memory.” This framework imposes a stricter discipline. Before attributing observed persistence to an internal mechanism, one must map the origin, relays, archives, secondary surfaces, comparisons, profiles, quotations, and the effective hierarchy of sources.
This is less a test of availability than an audit of the survival of a framing.
What this framework covers
- cases where a deleted, corrected, retracted, or 404 page keeps influencing outputs;
- cases where a secondary source becomes more prescriptive than the current source;
- cases where several regimes are collapsed under the word “memory”;
- production of an actionable diagnosis for endogenous, exogenous, or mixed correction.
What this framework does not cover
- internal forensic analysis of a closed model;
- strict security incidents;
- definitive proof of what a model saw during training;
- purely local session corrections when the problem remains fully stateful and involves no external relay.
Operational definition
Here, interpretive persistence after deletion designates the situation in which a framing remains mobilizable or dominant inside AI outputs even though the original source has been deleted, corrected, deindexed, retracted, replaced, or made unavailable.
The point of the audit is not merely to observe persistence. It is to identify through which regime the persistence is maintained.
Audit triggers
Launch this framework when at least one of the following signals appears:
- an answer still cites, directly or indirectly, a disappeared page;
- a third-party source keeps reformulating a verdict that has become false or obsolete;
- a ranking, profile, or comparison dominates the official source despite correction;
- a 404 is observed, yet the framing remains unchanged across several outputs;
- local correction seems ineffective or produces only a short-lived improvement.
The five regimes to distinguish before any conclusion
1. Still-active origin
The supposedly deleted source remains accessible in practice, cacheable, served differently, repeated under another URL, or available through an equivalent version.
2. Citation persistence
The origin has become secondary or unavailable, but reprises, profiles, lists, comparisons, or quotations continue to carry its framing.
3. Surviving authority
A secondary relay does more than repeat. It becomes, in practice, the source that frames reconstruction.
4. Interpretive remanence
An old interpretation returns intermittently, sometimes even after visible correction, because it remains easier to reconstruct than the current canon.
5. Neighborhood contamination
The framing comes neither directly from the origin nor from an identifiable quotation, but from a nearby semantic environment where the same story remains dominant.
A good audit may conclude that a case combines several regimes at once. The goal is not theoretical purity. The goal is the right remediation.
Framework rules
AP-1: no internal hypothesis without external mapping
As long as relays and secondary surfaces have not been qualified, the thesis of deep memory remains premature.
AP-2: the origin is not always the center of gravity
Once the framing has been redistributed, correcting the origin alone may become marginal.
AP-3: quotation is not the only proof of persistence
An answer may reproduce an older framing without explicitly citing the original source.
AP-4: a screenshot is not a diagnosis
Every observation must be replayed across a battery of formulations, surfaces, and temporal windows.
AP-5: real authority is measured by precedence, not supposed prestige
The key question is not “which source should be authoritative?” but “which source actually frames the answer today?”
AP-6: a 404 moves the investigation; it does not close it
Deleting the origin requalifies the problem. It does not necessarily solve it.
AP-7: the audit must produce a correction decision
If it does not end in a remediation hierarchy, the audit remains merely descriptive.
AP-8: every critical case must be replayed after correction
Without post-correction validation, the observed improvement may be only local noise.
Nine-step audit protocol
Step 1: formulate the testable allegation
Turn the case into a verifiable statement.
Example: “Despite the deletion of page X, the system still associates entity Y with status Z.”
Step 2: fix the canon and the date of truth
Document the current version, the date of correction, the authoritative pages, and explicit exclusions.
Step 3: verify the real state of the origin
Test the effective availability of the URL, its variants, redirections, caches, exports, duplicates, and equivalents.
Step 4: map the relays
List the surfaces that repeat or reinterpret the framing:
- rankings;
- profiles;
- directories;
- comparisons;
- media;
- PDFs;
- archives;
- local listings;
- derivative pages.
Step 5: qualify the status and strength of each surface
For each relay, qualify:
- current or historical status;
- informational density;
- proximity to the query;
- repetition across surfaces;
- ease of compression into an answer.
Step 6: execute a battery of prompts
Test multiple formulations, precision levels, comparison angles, nearby entities, and requests for justification.
The point is not to obtain one convincing output. The point is to see what resists when conditions vary.
Step 7: capture useful traces
Preserve citations, reformulations, displayed sources, missing proof, contradictions, and category shifts.
Step 8: classify the case
Assign a primary diagnosis and, if needed, secondary diagnoses: citation persistence, surviving authority, remanence, neighborhood contamination, or still-active origin.
Step 9: prescribe remediation
Establish a clear order:
- origin correction if necessary;
- correction of dominant relays;
- restoration of canonical precedence;
- publication of proof or clarification surfaces;
- re-tests and long-term monitoring.
Minimum qualification grid
A valid audit should at least score five dimensions qualitatively.
Relay intensity
Weak, medium, strong, or systemic according to the number and quality of the surfaces that prolong the framing.
Residual precedence
The degree to which a historical or third-party surface still frames before the current source.
Canon-output gap
Gap between the currently authorized version and the effective reconstruction observed.
Cost of correction
Endogenous only, targeted exogenous, extended exogenous, or hybrid correction with versioning, clarifications, and proof.
Associated risk
Reputation, identity, compliance, offer, territory, security, or agentic decision risk.
Remediation decisions according to diagnosis
If the origin remains active
Correct the origin, remove duplicates, clean redirections, and document the current state.
If the case is mostly citation persistence
Correct the dominant reprises, profiles, directories, lists, and comparisons that prolong the old framing.
If the case is mostly surviving authority
Explicitly requalify the source hierarchy and publish current surfaces that are easier to mobilize than the old relay.
If the case is mostly interpretive remanence
Reinforce canonical repetition, clarifications, cross-page coherence, and stability of critical formulations.
If the case is mostly neighborhood contamination
Reduce competing or ambiguous signals, work on disambiguation, and tighten the proximity graph.
The protocol that follows the audit depends on the dominant regime. When a third-party or historical surface still frames as if it remained first, the logical next step is the Protocol for exogenous deactivation of residual authority.
Expected deliverables
- Case register: allegation, date, criticality, evidence.
- Surface map: origin, relays, archives, neighbors, observed precedence.
- Regime qualification: primary diagnosis and secondary diagnoses.
- Remediation plan: endogenous correction, exogenous correction, proof to publish, order of treatment.
- Re-test report: post-correction results and resurgence indicators.
What a well-conducted audit helps avoid
- attributing a case too quickly to opaque “memory”;
- correcting only the initial page while the relay dominates elsewhere;
- mistaking residual visibility for a structural regime;
- multiplying ad hoc fixes without restoring canonical precedence;
- publishing impressionistic diagnoses that cannot be replayed.
Associated pages
- Black Hat GEO as symptom, not as a regime
- 404, deletion, and AI citation: what are we actually talking about?
- Deleted Wikipedia page: can it still act?
- Citation persistence
- Surviving authority
- Interpretive correction governance
- Protocol for exogenous deactivation of residual authority
- Exogenous governance: external graph stabilization