◎ OS PUB Apache 2.0 ← All specifications

P115 — AIEP — Cross-Participant Reasoning Style Compatibility Score

Publication Date: 2026-03-07 Status: Open Source Prior Art Disclosure Licence: Apache License 2.0 Author/Organisation: Phatfella Ltd Schema: AIEP_OS_SPEC_TEMPLATE v1.0.1 — https://aiep.dev/schemas/aiep-os-spec-template/v1.0.1

Framework Context

[0001] This disclosure operates within an AIEP environment as defined in GB2519711.2. It extends the Reasoning Style Fingerprint Substrate (P96) into the multi-participant session context defined by P110.

Field of the Disclosure

[0002] This disclosure concerns a deterministic mechanism for computing a ParticipantCompatibilityScore between any two session participants from their ReasoningStyleFingerprints (P96), and a SessionDiversityScore for the complete participant set, providing the session host with constitutional information about the structural reasoning diversity of the assembled group.

Summary of the Disclosure

[0003] A ParticipantCompatibilityScore is computed between any two participants A and B as a deterministic function of the structural divergence across their ReasoningStyleFingerprint fields (P96):

ParticipantCompatibilityScore(A, B) = 1 - NormalisedFingerprintDivergence(FingerprintA, FingerprintB)

where NormalisedFingerprintDivergence is computed as the mean normalised absolute difference across the six fingerprint fields: IncompatibilitySustainmentProfile, EvidenceWeightDispersionProfile, BranchDepthProfile, ArchivalReturnProfile, ConvictionVelocityProfile, and DivergenceTolerance.

[0004] A score of 1.0 indicates structurally identical reasoning styles. A score of 0.0 indicates maximum structural divergence.

[0005] A SessionDiversityScore is computed across the complete participant set as:

SessionDiversityScore = mean(NormalisedFingerprintDivergence(A, B)) for all pairs (A, B) in the session

[0006] A SessionDiversityRecord is generated comprising: SessionId; all ParticipantCompatibilityScore pairs; SessionDiversityScore; and SessionDiversityHash = H(CanonicalSerialise(CompatibilityScorePairs) ‖ SessionDiversityScore ‖ SchemaVersionId).

[0007] The SessionDiversityRecord is presented to the session host as constitutional information — not as a recommendation. The host is not directed to add or remove participants. They are shown the structural reasoning diversity of the assembled group as a substrate property.

[0008] A low SessionDiversityScore indicates participants with structurally similar reasoning styles — likely fast convergence, potential shared blind spots. A high SessionDiversityScore indicates structurally divergent reasoning — likely productive dissent, higher P97 sub-swarm probability.

[0009] Neither score influences session consensus computation, branch weighting, or P94 surfacing behaviour. They are informational substrate outputs only.

Licence

Abstract

A cross-participant reasoning style compatibility scoring mechanism is disclosed extending P96 ReasoningStyleFingerprint into the P110 session swarm context. A ParticipantCompatibilityScore is computed between each pair of session participants as a deterministic function of the structural divergence across their six-field ReasoningStyleFingerprints. A SessionDiversityScore aggregates pairwise divergence across the complete participant set. Both scores are presented to the session host as constitutional information about the structural reasoning diversity of the assembled group. Neither score influences session consensus, branch weighting, or surfacing behaviour.