Morphogenesis Chamber · reproducible scenarios

Change the input. Watch the pathway.

Each scenario below has a fixed initial state, a controlled perturbation and a deterministic outcome. Every outcome carries the rule that fired and the invariants that preserved identity. All scenarios are inspectable in text without JavaScript.

Disposition, per scenario: controlled-simulation = deterministic in-page data ·live-inspectable = pulls from a real edge endpoint (K23 KV / R2) ·research-hypothesis = asserted but not yet demonstrated. Every card labels its own disposition.

1. scenario-0controlled-simulation

   A first look — one entity, one relation, one piece of evidence

   The smallest complete demonstration: a single entity with a single admitted relation, permitted by a single stated rule.

   Initial state

   • e1 — Entity A entity
   • e2 — Entity B entity
   • c1 — Claim: A relates to B claim
   • v1 — Evidence source S1 evidence

   Relations

   • v1 → c1 (admits)
   • c1 → e1 (contextualises)
   • c1 → e2 (contextualises)

   Perturbation

   add-corroborating-evidence

   A second evidence source is admitted for the same claim.

   Added entities

   • v2 — Evidence source S2

   Added relations

   • v2 → c1 (admits)

   Outcome

   Verdict: permitted

   Rule that fired: inv-provenance-continuity

   The claim now has two admitting sources. Its provenance continuity is enhanced — both paths back to origin remain reachable.

   What did not change (Invariant Lens)

   • Entity A identity
   • Entity B identity
   • Original admission by S1
   • Claim identifier

   Applicable invariants

   • inv-provenance-continuity — Provenance of every admitted claim remains reachable

   Provenance

   • geometry-intelligence-foundations — §2

   Limitations

   • This scenario is a controlled simulation on fixed data; no external system is queried
   • Corroboration does not make the claim true — it makes its admission better supported

   Machine-readable definition: /api/scenarios/scenario-0.json

2. scenario-1controlled-simulation

   Corroborating evidence propagates through a small relational geometry

   Adding corroborating evidence to a peripheral claim increases the confidence of downstream claims that depend on it, all within the same authority boundary.

   Initial state

   • e1 — Author entity
   • e2 — Publication entity
   • e3 — Framework entity
   • c1 — Author wrote Publication claim
   • c2 — Publication defines Framework claim
   • v1 — ORCID record evidence

   Relations

   • v1 → c1 (admits)
   • c1 → e1 (contextualises)
   • c1 → e2 (contextualises)
   • c2 → e2 (contextualises)
   • c2 → e3 (contextualises)

   Perturbation

   add-corroborating-evidence

   A DOI record is admitted, corroborating both the authorship and the framework definition.

   Added entities

   • v2 — DOI record

   Added relations

   • v2 → c1 (admits)
   • v2 → c2 (admits)

   Outcome

   Verdict: permitted

   Rule that fired: inv-provenance-continuity

   Both claims now have two independent admitting sources. The relational geometry has not restructured — corroboration is not restructuring.

   What did not change (Invariant Lens)

   • Entity identities
   • Edge topology
   • Original ORCID admission path
   • Claim identifiers

   Applicable invariants

   • inv-provenance-continuity — Provenance of every admitted claim remains reachable

   Provenance

   • governed-relational-geometry — §3

   Limitations

   • Confidence propagation model is illustrative; real institutional systems use richer confidence semantics
   • Corroboration here means "another admitting source", not "external verification"

   Machine-readable definition: /api/scenarios/scenario-1.json

3. scenario-2controlled-simulation

   A context change preserves identity while altering permitted transformations

   The evaluation context is changed; the entity identity is preserved by the identity invariant, and the set of permitted transformations changes because the applicable authority boundary changed with the context.

   Initial state

   • e1 — Claim record entity
   • ctx1 — Context: research context
   • v1 — Evidence pack evidence
   • c1 — Admissible under research context claim

   Relations

   • v1 → c1 (admits)
   • ctx1 → c1 (contextualises)
   • c1 → e1 (contextualises)

   Perturbation

   change-context

   The evaluation context changes from "research" to "institutional assurance". The entity is the same; the invariants that apply are different.

   Added entities

   • ctx2 — Context: assurance

   Added relations

   • ctx2 → c1 (contextualises)

   Outcome

   Verdict: revised

   Rule that fired: inv-assurance-capability-separation

   The claim record entity persists unchanged (identity invariant). Its status is revised from admitted to capability-claim-pending-assurance, because the assurance context requires an evidence trail the research context did not.

   What did not change (Invariant Lens)

   • Claim record entity identity
   • Evidence pack contents
   • Research-context admission still valid within research context

   Applicable invariants

   • inv-identity-preservation — Identity is preserved under context change
   • inv-assurance-capability-separation — Capability claims do not cross into assurance claims without evidence

   Provenance

   • machine-entity-comprehension — §3
   • evidence-governed-intelligence — §2

   Limitations

   • Context is modelled as a single discrete node; real contexts are richer structures
   • The scenario shows that context change is informative; it does not certify any real system

   Machine-readable definition: /api/scenarios/scenario-2.json

4. scenario-3controlled-simulation

   Conflicting evidence blocks a pathway; the block itself is recorded

   Two admitting sources make contradictory claims about the same entity. The transformation that would admit the second is blocked; the block, the rule, and both admissions remain visible.

   Initial state

   • e1 — Entity X entity
   • c1 — X has property P claim
   • v1 — Source S1 evidence

   Relations

   • v1 → c1 (admits)
   • c1 → e1 (contextualises)

   Perturbation

   introduce-conflicting-evidence

   A second source asserts that X does NOT have property P. Both sources are equally weighted.

   Added entities

   • c2 — X does NOT have property P
   • v2 — Source S2

   Added relations

   • v2 → c2 (admits)
   • c2 → c1 (contradicts)

   Outcome

   Verdict: blocked

   Rule that fired: inv-contradiction-blocks

   The pathway that would collapse c1 and c2 to a single verdict is blocked. Both claims remain visible with their evidence; the contradiction and the rule that blocked resolution are recorded. Adjudication is deferred to a stated authority.

   What did not change (Invariant Lens)

   • Both admissions
   • Both sources
   • Entity X identity
   • The block itself is recorded (not hidden)

   Applicable invariants

   • inv-contradiction-blocks — Contradictory evidence blocks the pathway that admits it
   • inv-authority-bounded-conclusions — Conclusions do not exceed declared authority

   Provenance

   • governed-relational-geometry — §4

   Limitations

   • The scenario assumes equal source weighting; real adjudication uses confidence and authority differentials
   • The block is a design choice — a different system might auto-resolve; here we surface the disagreement

   Machine-readable definition: /api/scenarios/scenario-3.json

5. scenario-4live-inspectable

   A federation update propagates a shared fact without merging authority

   NODE-1 admits a new claim. NODE-23 receives the shared fact under the federation invariant that preserves per-node authority. NODE-23 admits the claim under its own rule of admission, or refuses it, without collapsing NODE-1's admission.

   Initial state

   • n1 — NODE-1 (ktsglobal.live) entity
   • n23 — NODE-23 (this node) entity
   • c1 — Claim admitted at NODE-1 claim
   • v1 — NODE-1 evidence surface evidence

   Relations

   • v1 → c1 (admits)
   • c1 → n1 (contextualises)

   Perturbation

   federation-propagation

   The federation shared-facts channel carries c1 to NODE-23. NODE-23 evaluates admission under its own rule.

   Added relations

   • c1 → n23 (transforms)

   Outcome

   Verdict: permitted

   Rule that fired: inv-federation-authority-preservation

   NODE-23 records that NODE-1 has admitted c1. NODE-23 has not by that fact alone admitted c1. NODE-1's admission remains sovereign; NODE-23's decision to admit or refuse is separate and recorded independently.

   What did not change (Invariant Lens)

   • NODE-1 authority
   • NODE-23 authority
   • Original admission provenance
   • Federation membership boundary (NODE-14 tombstone respected, NODE-1 sacred surface respected)

   Applicable invariants

   • inv-federation-authority-preservation — Federation membership does not transfer authority between nodes
   • inv-node-1-sacred — NODE-1 surface is sacred — only additive registry mentions permitted
   • inv-node-14-tombstone — NODE-14 is a tombstone precedent — no additive mentions, no modification
   • inv-provenance-continuity — Provenance of every admitted claim remains reachable

   Provenance

   • sovereign-web4-federation — §4

   Limitations

   • This scenario reads live federation state via the /api/federation/node.json surface; the outcome depends on the actual K23 KV seed populated after Fed Stage 6
   • Until Fed Stage 6 completes, the scenario renders in "pending K23 seed" mode with the pathway shown but no live counter

   Machine-readable definition: /api/scenarios/scenario-4.json