Claim Register — Post-Formalization Audit
Scoped to claims touching: Δh, Δc, detachment, rollback, closure,
sink/attractor language, terminal families, and “long enough.”
Generated 2026-04-03 after static topology and persistence model results.
Status key
- BROKEN — Lean directly falsified or the claim conflates proven-distinct concepts
- STALE — Not wrong, but uses framing that tonight’s work showed is imprecise
- SOUND — Survives formalization; correctly framed
- OPEN — Not yet testable; needs future work
Register
1. “Δh is the universal sink”
| Field |
Value |
| Location |
working/cybernetic-failure-taxonomy/taxonomy-relationships.md:136 |
| Claim |
“Δh is the universal sink — any failure that persists long enough becomes hysteresis” |
| Bucket |
Was presented as structural; is actually temporal |
| Status |
BROKEN |
| Tool |
Lean (TaxonomyGraph.lean — closure classification) |
| Fix |
Static topology yields three terminal families {Δg,Δa}, {Δx}, {Δh}. Δs and Δk cannot reach Δh. Universalization of Δh must be stated as a temporal hypothesis, not a graph fact. |
2. Δh property-based clarification (contradicts #1)
| Field |
Value |
| Location |
working/cybernetic-failure-taxonomy/taxonomy-role-map.md:72-75 |
| Claim |
“Δh as ‘universal sink’ means it’s the destination of uncorrected failures, not that every old failure is hysteretic. Discrimination criterion: pathological when persistence mechanism is self-referential rather than environment-referential.” |
| Bucket |
Structural (property-based definition) |
| Status |
SOUND — but buried under the broken version in #1 |
| Tool |
n/a |
| Fix |
Promote this as the canonical definition. #1’s temporal framing should be replaced with this property-based one. |
3. “Prolonged detachment leads to reset failure”
| Field |
Value |
| Location |
Implicit across taxonomy prose; explicit in persistence model design |
| Claim |
Sustained contiguous detachment is what produces hysteresis |
| Bucket |
Was temporal-spec; partly falsified |
| Status |
BROKEN (as necessary condition) |
| Tool |
Lean (PersistenceModel.lean — invariant 5: hysteresis_without_warn) |
| Fix |
Reset failure is driven by cumulative rollback depletion under detached commits. Prolonged contiguous detachment is sufficient but not necessary. Repeated short episodes suffice. |
| Field |
Value |
| Location |
preprint/06-temporal-closure-requirements/SI-C_Theory_Comparison.md:63-68 |
| Claim |
“Meta-representation requires temporal persistence. A higher-order thought about a first-order state needs both to persist long enough for the relationship to exist.” |
| Bucket |
Temporal-spec |
| Status |
STALE |
| Tool |
Prose revision |
| Fix |
Reframe as coexistence constraint: both states must be simultaneously available in the representational substrate. “Long enough” is implementation, not the theoretical claim. |
5. Attractor basin geometry (Second Law)
| Field |
Value |
| Location |
preprint/02-second-law-organizations/second_law.md:242-293 |
| Claim |
Basin A (high fidelity) is narrow, Basin B (low fidelity) is broad. Stochastic transitions preferentially move A→B because Ω_B » Ω_A. Return probability vanishes. |
| Bucket |
Structural (phase-space geometry) |
| Status |
SOUND — the asymmetry is geometric, not temporal |
| Tool |
n/a |
| Fix |
None needed for the core claim. Minor: “settling into” (line 281) should say “accessible equilibrium set under constraints” to avoid temporal implication. |
6. “Long Quiet” phenomenological signature
| Field |
Value |
| Location |
preprint/02-second-law-organizations/second_law.md:771-782 |
| Claim |
“Everything seems fine while Δt increases (metastable in Basin A under increasing effective heat)” |
| Bucket |
Temporal-spec |
| Status |
SOUND — correctly framed as temporal phenomenon from static cause |
| Tool |
n/a |
| Fix |
None. |
| Field |
Value |
| Location |
docs/method/falsification.md:53-62 |
| Claim |
Metastable requires: not in equilibrium, maintained by named active buffer, identifiable domain-break condition. “If you cannot name the buffer, the serialization mechanism, and the domain-break condition, you do not get to use the word.” |
| Bucket |
Structural (property-based definition) |
| Status |
SOUND |
| Tool |
n/a |
| Fix |
None. This is the correct framing. Cross-reference from Δh claims to reinforce. |
8. Δh lateral effects (“normalizing what should be temporary”)
| Field |
Value |
| Location |
working/cybernetic-failure-taxonomy/taxonomy-structured-pass.md:194-198 |
| Claim |
“Δh doesn’t generate new failure types so much as it locks in existing failures and then erodes the ability to recognize them.” Role: “Universal sink + lateral generator.” |
| Bucket |
Ambiguous — mixes state description with causal process |
| Status |
STALE |
| Tool |
Prose revision |
| Fix |
Distinguish: (a) Δh as a state (self-referential persistence) vs. (b) Δh→Δn, Δh→Δc as lateral reinforcement (already modeled separately as reinforces relation in Lean). Drop “universal sink” label. |
9. “Designed never to remain itself long enough”
| Field |
Value |
| Location |
working/claimant-transition-addendum.md:211-217 |
| Claim |
“It is not a worker because we designed it never to remain itself long enough to complain.” |
| Bucket |
Normative (governance/rights framing) |
| Status |
STALE |
| Tool |
Prose revision |
| Fix |
“Long enough” obscures the structural claim: the system lacks the invariants necessary for persistent identity. Reframe as “designed to lack the structural conditions for self-continuity.” |
10. Capacity-constrained stability (Paper 9)
| Field |
Value |
| Location |
preprint/09-capacity-constrained-stability/capacity_constrained_stability_complete_paper.md:279-285 |
| Claim |
“When shock arrival rate during the response window exceeds processable volume, the institution cannot maintain function long enough to implement adaptive responses.” |
| Bucket |
Temporal-spec (temporal outcome from static cause) |
| Status |
SOUND — correctly separates static constraint from temporal consequence |
| Tool |
n/a |
| Fix |
None. |
11. P25 §5 algebraic adjudication: aggregation does not rotate the observability subspace
| Field |
Value |
| Location |
preprint/25-epistemic-border-control/epistemic_border_control.md, §5 (“Algebraic adjudication” subsection through the closing core line) |
| Claim |
“Aggregation improves SNR; it does not rotate the observability subspace.” Stacked-witness observability matrix preserves the kernel; least-observable subspace is invariant under homogeneous replication. Paper 24’s clean aggregation is therefore not sufficient for substitution-freedom. |
| Bucket |
Structural (linear-algebra adjudication of a sibling-vs-nested decision) |
| Status |
SOUND — kernel preservation and Gramian scaling both proven |
| Tool |
Lean (Paper25EpistemicBorderControl.lean — ker_replicateRows_eq_ker, replicateRows_transpose_mul) |
| Fix |
Subspace-vs-vector precision added 2026-05-03 as a clarifying paragraph in §5: when the smallest singular value is degenerate, the invariant is the unobservable subspace, not a privileged $v_\text{min}$ vector. Explicit Gramian identity $(\mathbf{1}_N \otimes O_T)^\top (\mathbf{1}_N \otimes O_T) = N \cdot O_T^\top O_T$ included. |
12. P25 §3.1 Theorem 1: observation-equivalent states get identical control sequences
| Field |
Value |
| Location |
preprint/25-epistemic-border-control/epistemic_border_control.md, §3.1 (“Theorem 1 (static observability-asymmetry substitution)”) |
| Claim |
“Any controller whose policy depends only on ${y_0, \ldots, y_{T-1}}$ assigns the same control action sequence to $x$ and $x’$” when the observation trajectories agree. The structural refusal: observation geometry forecloses target regulation regardless of controller sincerity. |
| Bucket |
Structural (epistemic-access lemma; the policy has no distinguishing input) |
| Status |
SOUND in its load-bearing core — observation-equivalence ⇒ policy-equivalence is rw [h] |
| Tool |
Lean (Paper25EpistemicBorderControl.lean — obsEquiv_policy_same, target_distinct_policy_same) |
| Fix |
None to the structural refusal claim. The paper’s prose proof additionally hand-waves a closed-loop induction (closed-loop observations track open-loop ones under common controller action). That induction is correct but is not the load-bearing claim; the structural refusal stands without it. The Lean theorem isolates the load-bearing core. The corollary target_distinct_policy_same carries target q x ≠ target q x' as an intentionally-unused hypothesis: the policy never sees the target, so target inequality cannot break policy equality. |
Summary
| Status |
Count |
Action |
| BROKEN |
2 |
Rewrite with corrected claims |
| STALE |
3 |
Tighten framing, remove temporal conflation |
| SOUND |
6 |
No change; some need cross-referencing |
| OPEN |
1 |
Δc→Δh dynamics (partially formalized, external repair next) |
Entries #1–#10 are from the original 2026-04-03 audit, scoped to claims touching Δh, Δc, detachment, rollback, closure, sink/attractor language, terminal families, and “long enough.” Entries #11–#12 (added 2026-05-03) cover Paper 25’s §5 sibling-vs-§N algebraic adjudication and §3.1 Theorem 1 epistemic-access core, formalized in LeanProofs/Paper25EpistemicBorderControl.lean.
Priority rewrites — DONE (2026-04-03)
- taxonomy-relationships.md:136 —
Kill “universal sink” framing DONE. Replaced with three-terminal-family result + pointer to property-based definition.
- taxonomy-structured-pass.md:198 + :262 —
Drop “universal sink” from Δh role description DONE. Changed to “persistence sink (one of three terminal families)” in both role description and classification summary.
- taxonomy-role-map.md:75 —
Promote buried good definition DONE. Added Lean formalization result, three terminal families, and clarified temporal vs. graph distinction. Kept the self-referential persistence criterion as canonical.
- Implicit “prolonged detachment” assumption — Addressed in PersistenceModel.lean invariant 5 and RESULTS memo. Prose in taxonomy files now references cumulative rollback depletion rather than contiguous duration.