Worked Example — CNS 8.0 Resolves a Conditional Contradiction
Input account A
SNO-A:
Claim A1: Treatment X reduces symptom Y.
Evidence: Study 1, Study 2.
Relation: Study 1 supports A1.
Relation: Study 2 supports A1.
Input account B
SNO-B:
Claim B1: Treatment X does not reduce symptom Y.
Evidence: Study 3, Study 4.
Relation: Study 3 supports B1.
Relation: Study 4 supports B1.
Step 1 — Evidential Entanglement
The evidence sets overlap through shared measurements and trial endpoints. Entanglement is moderate/high.
Step 2 — Chirality
The accounts disagree over the same predicate:
reduces(X,Y)
Evidence-polarity chirality is high because the same endpoint is interpreted in opposite directions.
Step 3 — Antagonist report
The Antagonist finds:
- different dosage ranges;
- different age subgroups;
- different measurement windows;
- no direct citation failure;
- contradiction persists under original predicate vocabulary.
Step 4 — Zero-temperature closure
Strict closure proves:
Study1 supports reduces(X,Y) under high_dose.
Study2 supports reduces(X,Y) under high_dose.
Study3 supports not_reduces(X,Y) under low_dose.
Study4 supports not_reduces(X,Y) under low_dose.
The original predicate reduces(X,Y) remains contradictory because dose context was missing.
Step 5 — Residual tensor
Residual mass concentrates around:
subject: Treatment X
predicate: reduces
object: Symptom Y
context: dose / subgroup
Step 6 — Predicate invention
Tensor factorization proposes:
latent predicate L1: high_dose_context
latent predicate L2: low_dose_context
Grounding critic finds dosage spans in evidence atoms. The predicates pass initial grounding.
Step 7 — Synthesized SNO
SNO-C:
Claim C1 strict: Treatment X reduces symptom Y in high-dose contexts supported by Study 1 and Study 2.
Claim C2 strict: Treatment X does not show reduction of symptom Y in low-dose contexts supported by Study 3 and Study 4.
Claim C3 likely: Dose context explains the apparent contradiction.
Residual: Subgroup interaction remains unresolved.
Step 8 — Orthesis loop
Render SNO-C to language, re-ground it, and compare logic state.
If:
G(S(T_C)) ≈ T_C
and proof traces remain intact, SNO-C becomes an orthesis candidate.
Audit report
The final report includes:
- proof traces for C1 and C2;
- latent predicate status for dose context;
- unresolved subgroup residual;
- possible worlds for subgroup interaction;
- confidence language.