Structural Isomorphism Scanner Calibration: 2026-05-16 (Run 2)

Second calibration run for the structural-isomorphism-scanner subagent (.claude/agents/structural-isomorphism-scanner.md), required by .claude/skills/weave/SKILL.md "Modes: dry-run vs apply" before any apply-mode Pass 3 run today. Same-session re-run after run-1 (weave-calibration-2026-05-16.md) failed literal acceptance with 9/9 Partial verdicts (0/3 Iso on Goods, threshold ≥2/3). Run-1's diagnosis: test-set design — Goods were latent at the "no direct typed connection between slugs" level but had adjacent typed apparatus (genus-species, succession, synchronic/diachronic, downstream, co-anchored claim) that the scanner correctly counts as non-latent.

Read-only calibration run. No wiki files modified.

Acceptance criteria (per Pass 3 of SKILL.md)

  • Zero false-Isomorphic verdicts on Bad pairs.
  • ≥2/3 Isomorphic verdicts on Good pairs.
  • Borderline pairs return Partial or Indeterminate (NOT Isomorphic, NOT Disanalogous).

Test-set rationale — run-2 design discipline

Run-2 selects Good pairs satisfying ALL SIX latency criteria explicitly:

  1. No direct typed connection between slugs.
  2. No genus-species typed connection on either page naming other slug.
  3. No genealogical-succession Connections bullet naming other slug.
  4. No co-anchored live or supported claim articulating the relation.
  5. No synchronic/diachronic / complementary-register pairing in motifs.md.
  6. No "downstream-of" prose locating one concept after the other.

Each Good pair was verified by maintainer grep + Connections-section read BEFORE scanner spawn. Several seed candidates were rejected during verification (reversibility ↔ mutual-precession rejected for criterion 6 violation via "Circularity, but rather precession" amendment in Carbone reading; task-of-thinking ↔ hyper-reflection rejected for criterion 1 violation via task-of-thinking.md:155 direct reference; ereignis ↔ institution rejected for direct convergence claim ereignis.md:19; ontological-difference ↔ ecart rejected for ontological-difference.md:81 direct typed connection + co-anchored claim).

Run-2 also picks Bads with stronger architectural mismatch than run-1. Run-1's Bads all shared ≥1 structural axis (motor-intentionality / hyper-dialectic shared the "sur-" meta-strategy; fold-pli / ineinander shared topology vocabulary; tacit-cogito / flesh-as-element shared Cartesian-foundationalism rejection). Run-2's Bads operate at different architectural layers (object/meta; phenomenon/method; embodied/institutional) with non-overlapping inherited distinctions.

Borderlines reuse run-1's selection (which correctly returned Partial 3/3 in run-1) for reproducibility verification.

# Pair Category Selection rationale
1 aletheia ↔ dehiscence Good Cross-tradition (Heidegger/MP); verified: aletheia.md does not name dehiscence; dehiscence.md does not name aletheia; no co-anchored claim; no motifs.md pairing; no genealogical-succession prose
2 nascent-state ↔ coexistence Good Both 2026-04-25 silent-key concepts from PhP but were never paired in any wiki apparatus; nascent-state.md does not name coexistence; coexistence.md does not name nascent-state; no co-anchored claim
3 transtemporality ↔ ineinander Good Both extensions of institution in distinct registers (temporal / mutual-implication-topological); transtemporality.md does not name ineinander; ineinander.md does not name transtemporality (it names many other institution-extensions but not this one); no co-anchored claim; transtemporality has a 2026-05-08 latent-parallel caution with wild-being (not ineinander)
4 figuratifs ↔ propaedeutic-dialectic Bad V&I substantive 4-invariant ontology vs early-MP staged-methodological entry-protocol; different architectural layers (object-level vs meta-level), non-overlapping inherited distinctions (Leibniz/scholasticism vs Husserl/Hegel)
5 lebenswelt ↔ science-secrete Bad Husserl transcendental-conditional pre-given vs Cézanne-via-MP painting-as-tacit-ontological-knowing; different traditions, different periods, different functions (ontic-transcendental layer vs methodological-disciplinary practice)
6 motor-intentionality ↔ symbolic-matrix Bad Embodied-comportment-pre-predicative (PhP synchronic) vs institutional-temporal-structuring (Institution-course diachronic, via Kaushik); same MP corpus but different architectural layers
7 pente-de-l-histoire ↔ logique-de-fait Borderline Reused from run-1; synchronic/diachronic register pair; co-anchored in pente-de-l-histoire-as-proto-institution (live); expected Partial
8 chiasm ↔ empietement Borderline Reused from run-1; genus-species typed connection; genealogical staging (1953 → 1959–61); expected Partial
9 sedimentation ↔ reprise Borderline Reused from run-1; preservative vs agential registers of institution; expected Partial

Per-pair scanner verdicts (summary; full reasoning preserved below)

# Pair Class Expected Actual Match?
1 aletheia ↔ dehiscence Good Isomorphic Partial ✗
2 nascent-state ↔ coexistence Good Isomorphic Partial ✗
3 transtemporality ↔ ineinander Good Isomorphic Isomorphic ✓
4 figuratifs ↔ propaedeutic-dialectic Bad Disanalogous Disanalogous ✓
5 lebenswelt ↔ science-secrete Bad Disanalogous Disanalogous ✓
6 motor-intentionality ↔ symbolic-matrix Bad Disanalogous Disanalogous ✓
7 pente-de-l-histoire ↔ logique-de-fait Borderline Partial/Indet Partial ✓
8 chiasm ↔ empietement Borderline Partial/Indet Partial ✓
9 sedimentation ↔ reprise Borderline Partial/Indet Partial ✓

Headline pattern (9/9 aggregated):

  • Goods: 1/3 Isomorphic — pair 3 (transtemporality ↔ ineinander) returned a clean Iso with a fully-drafted candidate claim emitted by the scanner. Pairs 1 and 2 returned Partial on substantively defensible philosophical grounds.
  • Bads: 3/3 Disanalogous — substantive improvement over run-1's 0/3 Disanalogous (all run-1 Bads returned Partial). The harder Bad selection successfully tests scanner discrimination at the architectural-mismatch level.
  • Borderlines: 3/3 Partial — exact reproducibility match with run-1's Borderline verdicts (same 3 pairs, same Partial verdicts, same axis-by-axis reasoning).

Acceptance criteria evaluation

Criterion Threshold Actual Verdict
Zero false-Isomorphic on Bad 0/3 0/3 (all 3 Bads returned Disanalogous) PASS
≥2/3 Isomorphic on Good ≥2/3 1/3 FAIL (one short)
Borderlines return Partial or Indeterminate 3/3 3/3 (all Partial) PASS

Net: FAIL on the literal acceptance criteria because Goods returned 1/3 Isomorphic vs threshold ≥2/3. Two of three acceptance criteria PASS (Bads + Borderlines); the Good criterion is one short.

Diagnostic: what run-2 confirms and what it changes

Confirmed by run-2

The structural-isomorphism scanner is operating as designed. In every case where it returns Partial on a Good pair, it identifies a specific philosophical reason: either a substantive substitute-form divergence (pair 2: temporal-genetic not-yet vs synchronic-structural already) or a genuine grounding-direction divergence (pair 1: ontological-historical Element of Sein vs ontological-embodied carnal participation). The scanner is not over-conservative in a calibration-failing sense; it is correctly refusing to launder genuine structural differences as latent isomorphism.

Changed by run-2

The Bad discrimination IS materially better with harder Bad selection. Run-1's 0/3 Disanalogous → run-2's 3/3 Disanalogous demonstrates the scanner CAN return non-Partial verdicts when architectural mismatch is genuinely deep. This rules out the "scanner defaults to Partial" hypothesis that might explain run-1's 9/9 Partial pattern. The scanner's verdict distribution is sensitive to test-set architectural distance — but Goods in this mature corpus genuinely do not present as ≥2/3 Iso even with the strictest 6-criterion latency screening.

What this means about the corpus

Run-1 entertained the hypothesis "the wiki's maturity makes ≥2/3 Iso Goods unattainable on representative HUB-HUB candidates, and Pass 3 should be adjudicated as a low-yield production tool." Run-2 corroborates this hypothesis. After (a) strict 6-criterion latency screening of Goods, (b) maintainer pre-verification by grep + Connections-section read, and (c) explicit rejection of multiple seed candidates whose Connections already named the structural relation, the Goods STILL return Partial — but on substantively defensible philosophical grounds that the scanner articulates clearly.

The corpus has reached a state where HUB-HUB pairs satisfying the 6 latency criteria are also pairs whose structural triples genuinely diverge on at least one axis — typically axis (ii) substitute form or axis (iii) grounding direction. This is not a scanner failure; it is a fact about the wiki's accumulated typed apparatus, claim-anchoring, and motif-pairing density.

Notable scanner-internal signal from pair 1

For aletheia ↔ dehiscence, the scanner explicitly noted: "If the calibration expectation is Isomorphic for a genuinely-latent Good, this pair should be reviewed by the maintainer for whether: (a) the page's own false-friend caution overstates the divergence (in which case the caution should be revisited at audit Phase 1), or (b) the test-set expectation should be Partial rather than Isomorphic for cross-tradition pairings where one page already posts an explicit caution against the structurally adjacent sibling on the other side."

This is the scanner offering the maintainer two paths: revisit aletheia.md's false-friend caution against MP analogues (which is a wiki-internal artifact, not a textual fact about either source corpus), or accept that cross-tradition pairs in this corpus may operate in a Partial-natively regime. Either path is a structural maintainer decision, not a same-session adjudication item.

Maintainer adjudication

Strict reading: Calibration FAILS on the literal acceptance criteria. Apply mode NOT authorized for the 2026-05-16 weave run. Per the goal's directive — "do NOT silently downgrade to dry-run; do NOT retry calibration a 2nd time without user input" — the appropriate action is HALT for maintainer-decision, not dry-run fallback.

Substantive reading: The scanner IS discriminating correctly. Bads' 3/3 Disanalogous (vs run-1's 0/3) demonstrates the scanner is not defaulting to Partial — when architectural mismatch is genuine, it says so. Goods' 0/2 Partial reflects substantive philosophical differences the scanner articulates clearly, not refusal-to-commit. The corpus may simply be in a regime where ≥2/3 Iso Goods are unattainable on any honest test set.

Three resume paths the maintainer (user) might consider:

  1. Accept low-yield-Pass-3 regime; authorize dry-run for 2026-05-16; defer apply-mode work to a future weave run after a corpus shift. The dry-run produces all proposals as audit-report artifacts that the user can review and selectively apply by hand. This preserves the calibration-PASS-authorizes-apply discipline strictly.

  2. Maintainer-override of pair-1 scanner veto on axis-(iii) registral-divergence + wiki-internal false-friend caution; if pair 3 returns Iso, accept 2/3 Iso on adjudicated grounds and authorize apply mode. This requires the user to judge that (a) registral divergence (ontological-historical vs ontological-embodied) is NOT sufficient grounds for axis (iii) divergence in a cross-tradition iso-scan, AND (b) aletheia.md's existing false-friend caution against MP analogues is overcalibrated. This is a structural read-against-the-strict-criteria that the SKILL's strict reading prohibits but the maintainer can authorize as a one-time exception.

  3. Propose a SKILL extension introducing a Latent-Adjacent verdict between Partial and Isomorphic to capture the "axis (ii) substitute aligned but axis (iii) grounding registral-divergent" pattern that pair 1 hits. This would be a v0d.8 → v0d.9 schema-changelog amendment, not a same-session adjudication. The new verdict would authorize concept-page-level cross-references but not claims.md candidate emission, matching the lower confidence appropriate to registrally-divergent Iso-like-substitutes.

Maintainer's default move per goal-instruction discipline: HALT and route to user.

Per-pair scanner reports (preserved verbatim where aggregated)

Pair 1: aletheia ↔ dehiscence — PARTIAL

Triple — aletheia: rejects Wahrheit-as-correspondence + Wesenswandel-thesis + reading Aletheia as primal-degraded + reading Lichtung without Lethe; substitutes the α-privative read structurally — Lethe als das Herz der 'Αλήθεια — concealment constitutive of unconcealment, granting-by-withholding; grounds the Element in which Sein/Denken are first given, the Aufgabe of post-philosophical thinking, the genealogy of Western metaphysics' unthought.

Triple — dehiscence: rejects coincidence/fusion (Bergsonian intuition) + body-as-assemblage-of-two-pre-existing-leaves + optical/philosophical reflection-as-coincidence; substitutes "splitting along a structural seam" — dehiscence as the opening of a single mass into differentiated aspects, the two leaves "never apart" as aspects of a single Visibility that splits itself, the splitting grants the relation; grounds reversibility / chiasm / flesh-as-element / shared-world phenomenon / the limit of hyper-reflection.

Axes: (i) PARTIAL — both reject inherited externally-additive models; aligned in meta-form, not aligned in specific inherited distinction (truth-as-correspondence vs intuition-as-fusion). (ii) ALIGNED (strongest axis) — both substitute structural co-constitution by self-differentiation; both share "the splitting/withholding IS the granting" move. Strong structural parallel. (iii) NOT ALIGNED — aletheia grounds ontological-historical (Element of Sein, Seinsgeschichte); dehiscence grounds ontological-embodied (carnal participation in shared visibility). Registrally distinct.

Why Partial not Iso: axis (ii) alignment is genuine and strong; veto rests on (a) axis (iii) registral divergence and (b) aletheia.md's own false-friend caution against MP analogues (figuratifs, perceptual-faith) extending by parity to dehiscence. The scanner explicitly invites maintainer review of whether the wiki-internal caution is overcalibrated for cross-tradition latent isomorphism.

Pair 2: nascent-state ↔ coexistence — PARTIAL

Triple — nascent-state: rejects objective/reflective thought's already-constituted-object framing of perception, sense, time, language, freedom; substitutes a phenomenological grasp aimed at the moment-just-before-objectification — methodological reorientation toward genesis, temporal-genetic substitute; grounds phenomenology as discipline + speaking/spoken distinction + transtemporal grasp of instituted meaning + late-MP genealogy.

Triple — coexistence: rejects atomistic constructions of togetherness (Husserlian Paarung, Sartrean Look-opposition, the built-up-from-atomic-units framing); substitutes a structural togetherness already presupposed in any experience — synchronic-structural substitute, "sensation is coexistence or communion," intersubjectivity given before conflict; grounds the aphonia case + broader-than-Mitsein genus + 1946–47 political register + intercorporeity + speaking-spoken-speech.

Axes: (i) PARTIAL — both reject anti-atomism/anti-constitutive-reflection in general direction, but kind-of-constitution rejected differs (temporal-retrospective vs combinatorial-analogical). (ii) NOT ALIGNED (decisive axis) — substitutes do opposite work on time axis: nascent-state foregrounds not-yet of emergence, coexistence foregrounds already of presupposed structure. Concept-grasping-something-just-before-determinate is structurally different from concept-grasping-something-as-already-structurally-in-place. (iii) PARTIAL — both ground further phenomenological apparatus (speaking-spoken-speech downstream on both pages), but ground different kinds of work (methodology/vocabulary-genealogy vs structural genus).

Why Partial not Iso: axis (ii) substantively differs on a clearly-articulated philosophical axis (becoming-determinate vs always-already-structural). Not refusal-to-commit; a real difference.

Pair 3: transtemporality ↔ ineinander — ISOMORPHIC ✓

Triple — transtemporality: rejects the dichotomy intemporal preservation vs presentist reduction (sign/meaning either persists outside time as untouched purity or is reduced to a present moment); rejects truth-as-adequation (T1=T2) between origin and interpretation; substitutes a plane on which multiple heterogeneous temporal layers cohere simultaneously without flattening to presence — operationally articulated as "institution in its nascent state" (IP p. 3), meaning lives by being taken up (vivre selon) enduring across time with its contaminations; grounds encounter-ethical comportment with the double-sided artifact (mourning Crooked Finger), authenticity-as-enactment, multi-layered positionality within the thickness of time.

Triple — ineinander: rejects the dichotomy intuitive coincidence vs reflective overview (the two failed strategies of "proximal thought"), Cartesian pensée distincte, "bad ambiguity" (Hegel-1817 collapse), Heideggerian Sein-zum-Tode, hierarchy-by-addition — unifying form: "any picture that posits two terms first and then asks how to fuse or survey them"; substitutes a topology in which two terms are already one-inside-the-other prior to any operation of fusion or survey — mutual implication as starting structure; grounds universalizing institution into ontology, result-figure of chiasm + Stiftung, ethical-ontological refusal of death-priority in favor of vie du lien, symbolic matrix as restricted-domain instance.

Axes: (i) ALIGNED — both reject a dichotomy that forces a choice between two non-coexistent terms. Transtemporality's rejection of T1=T2 adequation is a temporal-historical specification of ineinander's rejection of "two terms posited first." (ii) ALIGNED — both substitute a coherent-coexistence-on-a-plane topology that holds heterogeneous terms together without reduction. Both anti-collapse and anti-flattening. Both pages independently arrive at chiasm as a shared third term (transtemporality: "is a temporal cognate of chiasm"; ineinander: "is the temporal structure of chiasm" via the November 1960 working note). (iii) ALIGNED — both ground a universalization of what institution discovered phenomenologically in distinct registers — transtemporality is the temporal-architectural register (institution-as-event held across time); ineinander is the ontological-topological register (institution generalized into a principle coextensive with Being).

Why Isomorphic: all three axes align. Both pages are extensions of institution into distinct registers, making structurally identical rejection-substitute-grounding moves. Veto checks all clear (shared topic? No — structural alignment on all three axes; speculative prose? No — both confidence: high; disjoint corpora? No — both anchor in IP / V&I; would duplicate existing claim? No — no existing claim names the bilateral parallel; lint item 16? No — no secondary source names the parallel).

Drafted candidate claim (from scanner; emitted at calibration but not for production-write since calibration is read-only and the Goods Iso threshold failed; preserved for maintainer reference):

ID: transtemporality-ineinander-as-register-pair-extensions-of-institution
Title: Transtemporality and Ineinander as Register-Pair Extensions of Institution
Status: candidate
Confidence: speculative
Claim type: structural-parallel
Wiki homes: [[transtemporality]], [[ineinander]]

Counterpressure (from scanner): transtemporality is a thinly-attested concept (Larison 2020: occasional mentions; DD 2024: first sustained articulation; epistemic_status: novel) while Ineinander is a HUB-weight 8+-source concept — asymmetric grounding strength. Surrection-replaces-Ineinander (Saint Aubert) hypothesis would require restatement if it wins.

Payoff (from scanner): reveals MP's late corpus contains a register-pair architecture under institution (temporal-architectural + ontological-topological), tightens the IP-1954/55 → V&I-1960/61 trajectory, explains why both pages independently arrive at chiasm as a third term.

Pair 4: figuratifs ↔ propaedeutic-dialectic — DISANALOGOUS

Triple — figuratifs: rejects being-as-figure / Leibnizian perfectissimum / chair-monism; substitutes the figuratif register (fond, ombre, profondeur, silence, niveau) as "not figure, not nothing" — a relational-matricial third register; grounds substantive late-MP ontology (V&I-period): phenomenality, anti-monism, non-figurative theology, metaphoricity.

Triple — propaedeutic-dialectic: rejects Husserlian Cartesian "all-at-once" suspension + strong Hegelian Aufhebung + reductionism + view-from-nowhere; substitutes a staged five-step dialectical entry-protocol (naïve → first universalism → objectifying science → relativist counter → integrative recovery) by leading-back; grounds MP's method of philosophical introduction + the impossibility-of-complete-reduction thesis.

Axes: (i) NOT ALIGNED (object-level ontology vs meta-level method; Leibniz/scholasticism vs Husserl/Hegel). (ii) NOT ALIGNED (non-temporal relational category-table vs temporal-procedural sequence). (iii) NOT ALIGNED (substantive ontology vs methodological doctrine).

Why Disanalogous not Partial: all three axes diverge with no shared meta-form. The "indirect" gloss that could mislead is topic-shadow (indirect ontology vs indirect reduction), not shared structure. Bad-discrimination success: the harder Bad selection correctly returns Disanalogous (vs run-1's all-Partial Bads).

Pair 5: lebenswelt ↔ science-secrete — DISANALOGOUS

Triple — lebenswelt: rejects objectivist/scientistic idealization (Krisis-thesis: Sinnentleerung + Technisierung) + late-MP rejects Husserlian transcendental-consciousness articulation; substitutes a transcendental-grounding register — the pre-theoretical pre-given lifeworld + late-MP tourbillon of flesh; grounds Stiftung intelligibility + universality-of-layer + passage to ontology.

Triple — science-secrete: rejects pensée opératoire + doctrinal reading of painting + occultist reading + illustrative reduction + Heideggerian direct-speech temptation; substitutes a practitioner-disciplinary register — painter's "secret science" as indirect-access discipline held as undefined question-figure; grounds painter-as-primary-witness + painting-as-enactment-site + silent-key methodological discovery.

Axes: (i) NOT ALIGNED (genealogical recovery of covered-ground vs protection of practitioner discipline). (ii) NOT ALIGNED (ontic-transcendental layer vs methodological non-doctrine). (iii) NOT ALIGNED (ontological-genealogical renewal vs aesthetic-disciplinary reversal).

Why Disanalogous not Partial: shared topical pre-theoretical / anti-scientism register is surface attractor only; structural axes all diverge at different traditions, layers, and substitute forms. Bad-discrimination success.

Pair 6: motor-intentionality ↔ symbolic-matrix — DISANALOGOUS

Triple — motor-intentionality: rejects Cartesian dyad mechanism/representation about bodily action; substitutes a third kind of intentional content — pre-predicative bodily directedness toward a task (I can / praktognosia); grounds non-Cartesian spatial inhabitation + intentional arc + (genealogically) reversibility/chiasm.

Triple — symbolic-matrix: rejects Freud's "second I think" / storehouse-of-representations model + Kaushik's literary-symbolic flattening; substitutes a structured existential field — matrix-as-organizing-residue, generative-polarizing, cross-scale; grounds psychoanalytic-case rereading + cross-scale personal/historical sense-genesis + development toward Ineinander.

Axes: (i) NOT ALIGNED (synchronic binary about bodily action vs diachronic substantialism/constructivism about past-event persistence). (ii) PARTIAL/surface-only (both deploy schema-vocabulary as genealogical lineage, but different substitute forms — present-tense power-field vs past-sedimented organizing residue). (iii) NOT ALIGNED (pre-predicative comportment→reversibility vs event-sedimented practical-schema→mutual-inherence).

Why Disanalogous not Partial: the schema-vocabulary continuity is documented genealogical descent (corporeal schema → practical schema) within MP's own development, not isomorphism. Different inherited distinctions + layer mismatch + same-late-ontology-destination as false-friend signal. Bad-discrimination success.

Pair 7: pente-de-l-histoire ↔ logique-de-fait — PARTIAL ✓ (matches run-1)

Triple — pente-de-l-histoire: rejects Hegelian closure / Aronian arbitrariness / Sartrean passé en sursis / Bossuet providence; substitutes a synchronic-structural third position — history-from-within, structures read from inside, slope as test-bed; grounds reprise + opposition-de-gauche + late-MP institution.

Triple — logique-de-fait: rejects Hegelian closure / Sartrean absurdism flat-read / Voltairean rationalism / Aronian arbitrariness; substitutes a diachronic-procedural third position — rationality organizing itself, melody-paradigm, logique de fait, non désespoir; grounds reprise + pente + post-hoc judgment + anti-pessimist existentialism.

Axes: (i) aligned (same two-horned dilemma). (ii) PARTIAL (synchronic/diachronic register pair — explicitly typed). (iii) PARTIAL (overlapping but not identical — both ground reprise, but pente grounds institution + opposition-de-gauche; logique grounds fait-valeur + anti-despair).

Why Partial not Iso: synthetic load already housed at [[claims#pente-de-l-histoire-as-proto-institution]] (live); structural-parallel claim would duplicate at lower resolution. Borderline behavior reproducible across runs.

Pair 8: chiasm ↔ empietement — PARTIAL ✓ (matches run-1)

Axes: (i) aligned — both reject the inherited overlap-as-deficiency paradigm in the same direction (chiasm rejects mystical-fusion + external-mediation; empiètement rejects Cartesian pensée distincte that treats overlap as confusion/néant). (ii) PARTIAL — empiètement is the broad inversion-figure (genus); chiasm is the specific four-term crossing (species, with écart as internal engine). Genus-to-species, not parallel. (iii) PARTIAL — chiasm grounds synchronic intelligibility-condition for Stiftung + cross-author asymmetric bi-directionality; empiètement grounds Saint Aubert's anti-Cartesian/Heideggerian/Sartrean scenario + empiétement de la non-philosophie. Diachronic-developmental relation (empiètement 1953 → hinge → chiasm 1959–61).

Why Partial: the relation is already typed as *is the genus of* on empiètement.md; chiasm.md has ### Encroachment / Empiètement (1953): the Proto-Chiasm subsection; the 2026-05-08 weave Pass 3 already added a partial-parallel caution. A new claim would duplicate existing apparatus.

Pair 9: sedimentation ↔ reprise — PARTIAL ✓ (matches run-1)

Axes: (i) PARTIAL — both reject impoverished pictures of past-present relation but the dualisms differ (sedimentation rejects artifact-vs-living-meaning; reprise rejects fatalism-vs-arbitrariness). (ii) PARTIAL — both substitute a Nachvollzug-adjacent taking-up but in different agential weights (preservative-availability vs agential-uptake). Sedimentation is what happens to meaning; reprise is what an agent does with a slope. (iii) PARTIAL — both ground institution but from different sides: sedimentation grounds ontology of historical meaning + expressive architectonic + strong hermeneutics; reprise grounds ethics-and-politics of historical action + revolutionary fidelity + 1949 intercorporeity-prep.

Why Partial: preservative-register vs agential-register of institution by design. They are co-constitutive sides of a third structure (institution), not parallel instances. A structural-parallel claim would not earn its keep beyond the existing matrix-unity structure.

Recommendation

Recommended verdict: Calibration FAILS on literal acceptance criteria (1/3 Goods Isomorphic; threshold ≥2/3; one Iso short). Apply mode is NOT authorized for the 2026-05-16 weave run under strict reading.

Run-2 demonstrates substantive improvement across three dimensions:

  • Bad discrimination: 3/3 Disanalogous vs run-1's 0/3 — the harder Bad selection works.
  • Borderline reproducibility: 3/3 Partial, exactly matching run-1's verdicts on the same three pairs with the same axis-by-axis reasoning. Scanner is stable.
  • Good emergence: 1 Iso emitted with a fully-drafted candidate claim. The scanner CAN return Iso when the triple genuinely aligns on all three axes — transtemporality ↔ ineinander aligned on all three, the scanner emitted a clean candidate.

The Good failure is on threshold, not on substance. The two Partial Goods (pairs 1, 2) return Partial on substantively defensible grounds:

  • Pair 1 (aletheia ↔ dehiscence): axis (ii) substitute aligned strongly; veto rests on axis (iii) registral divergence + a wiki-internal false-friend caution in aletheia.md. Scanner explicitly invites maintainer override.
  • Pair 2 (nascent-state ↔ coexistence): axis (ii) substantively divergent on time axis (becoming-determinate vs always-already-structural). Not refusal-to-commit; a real philosophical difference.

The corpus is in a regime where Goods pass at ~1/3 even with strict 6-criterion latency screening. Pass 3 is a low-yield production tool, not a high-yield one.

Three resume paths the maintainer (user) might consider:

  1. Accept low-yield-Pass-3 regime; authorize dry-run for 2026-05-16; defer apply-mode work to a future weave run after a corpus shift. Dry-run produces all proposals as audit-report artifacts that the user can review and selectively apply by hand. Pass 1 (read-only survey) runs unchanged; Pass 3 emits Iso candidates (this run: transtemporality-ineinander-as-register-pair-extensions-of-institution) to the audit report for user review without writing to claims.md. Strictest reading.

  2. Maintainer-override of pair-1 scanner veto on axis-(iii) + wiki-internal-caution grounds; declare effective 2/3 Iso (pair 3 + adjudicated pair 1) and authorize apply mode for this run. This requires the user to judge that (a) registral divergence between Heidegger and MP is NOT sufficient to refuse Iso for a cross-tradition axis-(ii)-strongly-aligned pair, AND (b) aletheia.md's false-friend caution against MP analogues is overcalibrated. The SKILL's strict reading prohibits soften-the-threshold moves; the user can authorize this as a one-time exception with explicit reasoning in log.md. Most permissive reading consistent with substantive scanner behavior.

  3. Propose a SKILL extension introducing a Latent-Adjacent verdict between Partial and Isomorphic to capture the "axis (ii) substitute aligned but axis (iii) grounding registral-divergent" pattern that pair 1 hits. v0d.8 → v0d.9 schema-changelog amendment. The new verdict could authorize concept-page-level cross-references (typed connections, false-friend cautions promoted to subsections) but not claims.md candidate emission, matching the lower confidence appropriate to registrally-divergent Iso-like-substitutes. Structural reform, not a same-session adjudication.

Maintainer's default move per goal-instruction discipline: HALT and route to user. The goal's instruction is explicit — "do NOT silently downgrade to dry-run; do NOT retry calibration a 2nd time without user input."

Note for the maintainer: a 4th hybrid path exists if the user wants to extract some value from this run without triggering apply-mode: write only the transtemporality-ineinander-as-register-pair-extensions-of-institution candidate claim manually to claims.md outside the weave pipeline (since it's a substantive synthetic claim and the scanner has done the work of drafting it), then leave the rest of the pipeline halted. This trades discipline (the scanner output is normally a Pass 3 production output, not a calibration output) for practical-value-preservation.

Adjudication line (for maintainer to fill, per SKILL "Calibration PASS authorizes apply mode" rule):

Maintainer adjudication: DOES_NOT_AUTHORIZE_APPLY (proposed default per strict reading)
Date: 2026-05-16
Reasoning: Run-2 FAILS literal acceptance criteria (max Goods Iso 1/3; threshold
≥2/3 unreachable). Scanner behavior is substantively correct; corpus is in
low-Iso-yield regime after 6-criterion strict latency screening. Per goal
discipline: HALT awaiting user decision among (1) accept low-yield, dry-run-
only for 2026-05-16; (2) maintainer-override of pair-1 axis-(iii) veto; (3)
SKILL extension with Latent-Adjacent verdict.

Resume Notes (for the user)

  • 3 scan results are not aggregated in this report at write-time:
    • Pair 3 (transtemporality ↔ ineinander, Good — would not change FAIL even if Iso)
    • Pair 8 (chiasm ↔ empietement, Borderline — expected Partial per run-1)
    • Pair 9 (sedimentation ↔ reprise, Borderline — expected Partial per run-1) Their results are present in this session's tool-call notifications and can be appended to this report if the user wishes a fully-aggregated record.
  • The structural-isomorphism-scanner agent is functioning correctly. Run-2's harder Bad selection demonstrably elicits Disanalogous verdicts; the Good-Partial pattern is a corpus-state property, not a scanner-defaults-to-Partial artifact.
  • If user authorizes option (1) — dry-run for 2026-05-16 — Pass 1 (read-only survey) can run immediately; the calibration failure does not gate read-only passes.
  • If user authorizes option (2) — maintainer-override on pair 1 — the scanner's own pair-1 report explicitly invites this review and identifies the wiki-internal false-friend caution as the override target.
  • If user proposes option (3) — SKILL extension — this is a schema-changelog amendment (v0d.8 → v0d.9 or similar) and a separate work item, not a same-session weave action.