Failure Modes¶
Failure modes describe recurrent structural collapses that make records harder to inspect, verify, transform, compare, or preserve under disagreement.
A structural collapse occurs when a record fuses things that AR requires to remain separable.
Authoritative failure-mode source files live in:
Failure-mode element types live in:
Failure modes are not exhaustive. They name recurring risks that profiles, claims, traits, mappings, transformation checks, and conformance checks can detect or help prevent.
Core Concepts¶
| Concept | Meaning |
|---|---|
| Failure Mode | A common way records become unreliable or hard to inspect. |
| Structural Collapse | A record design failure where separate structures are fused. |
| Collapse Pattern | A recurring pattern where a record mixes things that should stay separate. |
| Collapse Detection | A claim, trait, mapping, check, or profile rule used to detect a collapse. |
| Collapse Remediation | Guidance for fixing or avoiding a record collapse. |
| Profile Failure Mode | A domain-specific failure mode defined by a profile. |
Core AR Failure Modes¶
AR defines six core structural-collapse patterns:
| Failure Mode | Collapse |
|---|---|
| Source vs. Interpretation | Source material is fused with an interpretation of what the source means. |
| Name vs. Identity | Identity is tied to a mutable name, title, address, label, or designation. |
| Content vs. Status | Content is mutated to encode current status. |
| Event vs. Record | A record describing an event is treated as the event itself. |
| Provenance vs. Authority | Provenance is treated as conferring authority. |
| Context vs. Claim | Contextual limits are folded into the claim itself. |
These are structural patterns, not final judgments about truth, authority, legitimacy, obligation, enforcement, or domain meaning.
Source vs. Interpretation¶
This collapse occurs when a record fuses source material with an interpretation of what the source means.
It may appear in:
- source records whose content field is an extracted thesis rather than an extracted passage;
- event records that re-characterize the event using interpretive vocabulary;
- domain records that paraphrase source material into an asserted conclusion.
Detection may use:
- source traceability checks;
- observation non-authority traits;
- interpretation non-mutation checks.
Remediation: keep source-bearing fields faithful to source material. Express interpretation in separate records, profile-specific interpretive records, claims, annotations, or cross-referenced structures.
Name vs. Identity¶
This collapse occurs when identity is tied to a name, title, address, section number, display label, or other mutable designation.
It may appear in:
- identity records keyed by current display name;
- place or asset records whose stable ID is a postal address;
- rule records identified only by the latest section number.
Detection may use:
- stable identifier checks;
- subject identity checks;
- identity-preserving transformation checks.
Remediation: use stable identifiers for records and subjects. Treat names, titles, addresses, display labels, section numbers, and other designations as mutable fields rather than identity.
Content vs. Status¶
This collapse occurs when content is mutated to encode current status.
It may appear in:
- rule records where repeal is represented by deleting rule text;
- policy records where supersession overwrites prior policy content;
- records that can no longer answer what the content said before the status changed.
Detection may use:
- content/status separation traits;
- status history checks;
- interpretation non-mutation checks.
Remediation: keep content separable from status. Express repeal, supersession, limitation, current status, or other status changes as separate status structure, status history, linked records, or declared transformations.
Event vs. Record¶
This collapse occurs when a record treats a representation of an event as the event itself.
It may appear in:
- event records rewritten when better information arrives;
- audit records where correcting a logged event mutates the original entry;
- records that cannot distinguish the occurrence from later descriptions.
Detection may use:
- subject mapping checks;
- transformation admissibility checks;
- record correction lineage checks.
Remediation: treat the event and the record describing the event as distinct. Corrections, additions, and reinterpretations should be new records or declared transformations that reference the original record.
Provenance vs. Authority¶
This collapse occurs when provenance is treated as conferring authority.
It may appear in:
- records that omit scrutiny because of producer reputation;
- profiles that conflate signed-by with true;
- verifiers that pass checks because provenance is trusted.
Detection may use:
- authority non-assertion traits;
- provenance traceability checks;
- claim outcome checks.
Remediation: record provenance as information about how the record came to be. Treat authority as a separate interpretive, profile-specific, institutional, or downstream claim rather than as something AR infers from provenance.
Context vs. Claim¶
This collapse occurs when contextual limits, scope, environment, observation conditions, or applicability boundaries are folded into the claim itself.
It may appear in:
- observation records that absorb caveats into their statements;
- rule records whose scope is silently incorporated into content;
- reports that treat partial outcomes as pass because context explains the failure.
Detection may use:
- scope declaration traits;
- context subject mappings;
- claim outcome vocabulary checks.
Remediation: keep context separate and citable. Claims may cite their context, scope, environment, observation conditions, or applicability boundary, but they should not absorb that context in a way that makes the claim appear stronger or more general than it is.
Detection¶
Failure modes may be detected by:
- claims;
- traits;
- field mappings;
- subject mappings;
- transformation checks;
- conformance checks;
- profile rules.
A detector reports a structural failure, risk, or finding.
It does not decide truth, authority, legitimacy, obligation, enforcement, or final domain meaning.
Remediation¶
Remediation guidance describes how records, profiles, mappings, or checks can be structured to avoid or repair a collapse.
Remediation does not rewrite records by itself.
Remediation does not decide final domain meaning.
Profile Failure Modes¶
Profiles may define additional domain-specific failure modes.
Profile failure modes compose with AR core failure modes. They may specialize AR failure modes, but they must not redefine AR failure modes or AR conformance outcomes.
A profile-defined failure mode should identify:
- the collapse;
- where it appears;
- how it is detected;
- relevant traits, claims, mappings, or transformation rules;
- remediation guidance.
Examples:
- a judicial-record profile might define holding-vs-dictum, citation-vs-dependency, or majority-claim-vs-court-endorsement collapses;
- a civic or institutional profile might define contribution-vs-influence or meeting-vs-decision collapses.
Why Failure Modes Matter¶
Many collapses are locally attractive because they make records simpler in the short term.
The cost appears later, when a record must be inspected, audited, transformed, mapped, compared, or reviewed under disagreement.
AR pays structural cost up front so downstream inspection is possible.
Boundary¶
Failure modes help AR preserve structural separability.
They do not decide:
- truth;
- authority;
- legitimacy;
- legal effect;
- enforcement;
- final domain meaning;
- whether a profile-specific interpretation is correct.