Research: Contract C non-deciding shadow RC0

[camerontjs-dot]

Classification

Draft Research Infrastructure / cross-repository Contract C representation evidence. Do not merge. No released Contract C mutation, canonical version assignment, Decision Engine production change, Contract E, release, tag, promotion, or authorization.

Exact lineage

• live Apparatus main at experiment start: c3563cff66d2c85dcbf575c693056e2d8e4563d4
• exact base branch: research/contract-c-non-deciding-shadow-rc0-base-20260910
• released Contract C schema blob held fixed: b0369de9b5c156322d6787261bbc7658a3b33781
• released Contract C validator blob held fixed: 9c75ccfbf2223578a8d1a7bf0c39673b394fbea4
• CAL repair-comparison PR #100 pressure-test head: aa5f0f1313e65e6d31493095214c76d743ca6d89
• frozen CAL four-candidate implementation blob: a7934b3c242dcf1c33a28121b3b141c9c6adc203
• CAL multiplicity run: 34506201889
• CAL multiplicity artifact: 10163899726, digest sha256:23ba4cb13870f0b2c5f213d3279faede9d17f19e638dea49444249328a0efb19

Question

Can the sole multiplicity-preserving survivor from CAL PR #100 be represented at the Apparatus boundary by adding only a non-polarized non_deciding contribution channel, while preserving every unrelated released Contract C 1.0 structural, canonicalization, reference-integrity, Contract-B-binding, causal-multiplicity, and whole-object-identity rule?

Frozen shadow delta

The research wire sentinel is deliberately non-canonical: research-non-deciding-rc0.

Relative to released Contract C 1.0.0, the shadow differs semantically at exactly two leaves:

1. properties.contract_c_version.const -> research-non-deciding-rc0;
2. $defs.contribution.properties.channel.enum -> [support, counterevidence, non_deciding].

No released schema or validator file was modified. The research validator delegates all unchanged structural semantics to the released Contract C validator through a compatibility copy and separately validates the real shadow object's research version/channel and identity.

Structural qualification

Run 34528850394: PASS.

• exact tested head: 361f962fb8b0a43b208d739fd57e9aa1f1b364a0
• artifact: 10172655148
• artifact digest: sha256:94b5edd27cc2515ce9af3c460d88a821c1d59d477c18c97d832a543df2479f24
• disposition: SUPPORTED_BOUNDED_TWO_LEAF_SHADOW_DELTA
• generated shadow result-set: result-set:b6c9a7aa7ab0f627fb4cb4d29eb7638a03e8e21f9f5c7aa360f42e99b357ee2b
• generated shadow SHA-256: sha256:34e3b84357e4424fb69c41fc228c2feda81d73d12bd103ee46132a9bb832f017

Observed positives/falsifiers passed:

• two distinct non-deciding causal contributions preserve independent_sufficient_alternatives;
• one causal non-deciding contribution plus one residual non-deciding contribution remains valid;
• legacy support/counterevidence remain valid in the shadow;
• wrong/missing Contract-B evidence references fail;
• unknown basis, causal cardinality errors, causal/residual overlap and unclassified retained contributions fail;
• arbitrary added fields and unknown channels fail;
• result-set and whole-object hash tampering fail;
• unrelated schema relaxation is detectable;
• canonical round trip is stable;
• released Contract C 1.0 rejects the shadow specifically on the research version sentinel and non_deciding channel;
• released Contract C regression stayed green: 17 passed, 8 skipped.

Exact committed handoff qualification

A static handoff packet was committed so the independent consumer receives specification/artifact bytes rather than this producer implementation.

Decisive handoff qualification run 34529151133: PASS.

• exact tested head: beb195c61e7a85bdb0cb350d85c03ca48177e23d
• artifact: 10172768614
• artifact digest: sha256:7af02c2230be29b893f3480ef25b82a0caf0ce334d3967997ac713ebb659170b
• handoff valid-shadow.json SHA-256: sha256:325962ebcdbf6af836bb6193a451524ccd40b4d10f2394ff9f703fbfce1ec1e3
• handoff result-set identity: result-set:4483272c4f6fbd9cb2362be7e3174bbd00aff3cf761d6c374897f3478818c9f0
• causal passages: u-a, u-b
• channel: non_deciding
• causal form: independent_sufficient_alternatives
• released Contract C regression: 17 passed, 8 skipped
• released 1.0 validator rejects the exact handoff bytes as intended.

Receipt commit: ad1ffbd7906a7cf34cce5afa906a5797cd4a14ff.

Bounded inference

Within this frozen Apparatus pressure test, the non_deciding representation survives the contract boundary without weakening unrelated Contract C 1.0 invariants. The smallest tested semantic schema delta is one added contribution-channel value plus a research-only wire-version sentinel.

This does not choose a production Contract C versioning strategy. It establishes only that the candidate is sufficiently coherent to hand to an independent downstream consumer.

Next falsifier

Give only the frozen handoff specification, schema delta, exact fixture and Contract-B index to a separate consumer. That consumer must independently recover both unresolved causal evidence references and their independent_sufficient_alternatives multiplicity, reject malformed reference/cardinality/version/hash mutations, and show that non_deciding is not silently reinterpreted as support, counterevidence, or operational authorization.

Not established

• canonical Contract C successor version;
• backward/forward compatibility policy;
• Decision Engine conformance;
• arbitrary semantic-family sufficiency;
• root / all_of CAL composition;
• Contract E / Authorization;
• merge, release, tag, or production promotion.