Exception Classification¶
Classification procedure — instantiates Metanarrative Coherence and Internal Consistency Check
Sorts each anomaly the narrative bumps into — legitimate boundary case, ambiguity needing qualification, repairable contradiction, or fatal disconfirmation — so counterexamples are neither absorbed ad hoc nor treated as automatic refutations.
Every broad narrative meets cases that don't quite fit, and the danger is not the anomaly itself but the undisciplined handling of it — quietly absorbing one counterexample as "the exception that proves the rule" while treating an identical one elsewhere as fatal. Exception Classification imposes a rule on that handling. It takes each anomaly the audit surfaces and sorts it into one of a fixed set of buckets — a legitimate boundary case the narrative was never meant to cover, an ambiguity that needs a qualifying clause, a repairable contradiction, or a genuine disconfirmation — and records why it landed where it did. Its defining move, the one that separates it from the mechanisms that find anomalies, is that it does not detect incompatibilities; it triages them, deciding what each anomaly means for the narrative and routing the disconfirming ones onward. It is the mechanism that stands directly against the archetype's twin failures: scope-creep absorption, where counterexamples are swallowed without revising the core, and overzealous contradiction-hunting, where every quirk is called incoherence.
Example¶
A literary critic has published a comprehensive reading of a modernist novel: the protagonist's every choice is driven by a buried guilt over his brother's death, and the whole book is that guilt working itself out. It's an elegant, totalizing interpretation. A graduate student stress-reading it collects the scenes that seem not to fit and runs them through an exception classification rather than declaring the reading wrong.
Scene by scene she sorts. The protagonist's cruelty to a stranger in chapter four seems to contradict a guilt-driven reading — but the narrator is established as unreliable there, so it goes in the ambiguity-needing-qualification bucket, not the refutation bucket. His indifference to his brother's grave is a genuine boundary case: the reading was always about psychological guilt, not ritual mourning, so it never claimed to cover that scene. But the final chapter, where he calmly explains he never felt responsible at all, lands squarely as a disconfirmation — the text asserts the opposite of the reading's core. Three anomalies, three different verdicts, each with its reason recorded. The critic's interpretation survives two of them intact and has to genuinely reckon with the third, instead of the anomalies being waved away or piled up into a false "the reading collapses."
How it works¶
- Collect the anomalies as raised. Take the cases the scan, trace, and traceability surfaced — plus any the narrative itself hedges around — as the items to be sorted; this mechanism classifies, it does not hunt.
- Apply a fixed sorting rubric. Route each anomaly through the same set of buckets — legitimate boundary, ambiguity needing qualification, repairable contradiction, fatal disconfirmation — asking the same questions in the same order, so identical anomalies get identical verdicts.
- Record the reason, not just the label. Note why each anomaly landed where it did, especially the test that separates a genuine boundary (the narrative never claimed this ground) from an ad hoc escape (the narrative is being narrowed only to survive).
- Route the outputs. Legitimate boundaries and qualifications go to the boundary register as accountable exceptions; disconfirmations are handed to the contradiction log as live tensions the audit must resolve.
Tuning parameters¶
- Bucket resolution — how many categories the rubric distinguishes. More buckets capture finer distinctions but slow the sort and invite disagreement over borderline cases; fewer buckets are fast but crush ambiguity and disconfirmation together.
- Ad-hoc bar — how much justification a "legitimate boundary" verdict requires. A high bar blocks scope-creep absorption but may reject genuine boundaries; a low bar lets any inconvenient case be relabeled "out of scope."
- Ambiguity tolerance — how readily an anomaly is parked as "needs qualification" versus forced to boundary-or-disconfirmation. High tolerance respects genuine plurality; too much of it becomes a holding pen where real refutations quietly wait forever.
- Consistency-of-application check — whether the rubric's verdicts are audited for even-handedness across anomalies. Enforced consistency is the whole point; skipping it lets the same kind of case be absorbed here and called fatal there.
- Escalation threshold — how many or how severe the disconfirmations must be before they trigger a verdict on the whole narrative rather than a local repair.
When it helps, and when it misleads¶
Its strength is even-handedness: by forcing every anomaly through the same rubric with a recorded reason, it stops the two opposite abuses at once — the story that absorbs every counterexample and the critic who treats every quirk as collapse. It preserves the archetype's hardest distinction, between legitimate complexity and hidden contradiction, by making "this is just a boundary case" a claim that must be justified rather than merely asserted. It is the guard against monster-barring — redefining the rule on the fly[1] so the awkward case is declared not to count.
Its failure mode is the rubric itself becoming the instrument of laundering: if the "legitimate boundary" bucket is easy to reach, classification gives ad hoc absorption a respectable-looking process, and every disconfirmation gets quietly reclassified as an out-of-scope edge. The classic misuse is exactly that — running the sort to manufacture a boundary for whichever anomaly threatens the conclusion. The discipline that guards it is the ad-hoc bar plus recorded reasons: a boundary verdict must rest on a scope the narrative stated before the anomaly appeared, not one drawn around the anomaly after the fact, and every verdict must survive being applied to the symmetric case.
How it implements the components¶
exception_and_boundary_register— its primary artifact: the record of each anomaly's verdict, separating accountable boundary cases and qualifications from the anomalies that turned out to disconfirm.contradiction_and_tension_log— the anomalies it classifies as genuine disconfirmations are written into the log as live tensions for the audit to resolve.
It does not define or apply the pairwise consistency_criteria that decide whether two claims clash in the first place — that detection is Contradiction Scan's; this mechanism sorts what such detectors surface. Nor does it draw the claim_dependency_map those anomalies are read against — that is Claim-Lattice Mapping's.
Related¶
- Instantiates: Metanarrative Coherence and Internal Consistency Check — it turns raw anomalies into accountable, evenly-judged exceptions.
- Consumes: Contradiction Scan supplies the flagged incompatibilities that this mechanism triages into boundary, ambiguity, or disconfirmation.
- Sibling mechanisms: Causal-Temporal Trace · Claim-Lattice Mapping · Contradiction Scan · Evidence-to-Claim Traceability · Red-Team Coherence Review · Revision Diff Review · Term-Stability Review · Counter-Narrative Probe · Scope-Boundary Stress Test
Editorial Notes¶
Form Classification¶
Form family: Decision, Gate & Allocation
Rationale: Exception Classification operates as a case-specific gate, selection, routing, prioritization, or resource disposition because it sorts each anomaly the narrative bumps into — legitimate boundary case, ambiguity needing qualification, repairable contradiction, or fatal disconfirmation — so counterexamples are neither absorbed ad hoc nor treated as automatic refutations.
Independent corroboration: The frozen evidence defines Exception Classification as 'Sorts each anomaly the narrative bumps into — legitimate boundary case, ambiguity needing qualification, repairable contradiction, or fatal disconfirmation — so counterexamples are neither absorbed ad hoc nor treated as automatic refutations', so its operative form is Decision, Gate & Allocation.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Philosophy
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Universal
Rationale: Classifying counterexamples as boundary cases, qualifications, contradictions, or disconfirmations arises from philosophical conceptual and theory appraisal.
Related originating lineages:
- Literature & Literary Theory — Narrative theory materially contributes the treatment of anomalous cases relative to a metanarrative's scope and coherence. The operational example and narrative-coherence use materially draw on interpretive criticism and close reading.
Review resolution: Both reviewers agree that philosophy is primary. I retain literature_literary_theory only as formative origin lineages; cross_disciplinary_synthesis is appropriate because the final form materially combines the agreed primary with the retained formative lineages. Reach is universal because the structure is portable across essentially any domain with the stated problem, an applicability judgment kept separate from provenance. Encyclopedia synthesis is true because the exact generalized artifact is an encyclopedia-authored combination or refinement. No unresolved historical ambiguity remains after reconciling the secondary fields.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; medium confidence.
References¶
[1] Lakatos, I.; edited by Worrall, J., and Zahar, E. Proofs and Refutations: The Logic of Mathematical Discovery. Cambridge University Press (1976). Defines monster-barring as an ad hoc redefinition that excludes an inconvenient counterexample from the relevant class. registry ↩