RFC: HELP — Human Epistemology Layer Protocol for plural-intelligence governance

[RobynAwesome]

RFC status

Candidate: HELP — Human Epistemology Layer Protocol
Expansion: Human Epistemology Layer Protocol
Origin: 2026-09-19 reality → cloud synthesis from KPGS policy/inclusion work
Implementation state: HOLD_PENDING_BGR
Authority boundary: This issue authorizes research and specification only. No RUNE/Algiz code, schema mutation, public-canon mutation, or issue closure is authorized until BGR / Baby Girl Rune reviews the proposal and explicitly promotes implementation.

---

Thesis

Current governance systems often treat "inclusion" as participation inside a model whose relevant categories, evidence types, and decision logic have already been fixed.

HELP proposes a stronger test:

Inclusion is not satisfied by placing more people inside an already-fixed model. Inclusion becomes operational when realities outside the model are able to challenge, alter, or expand the model before consequential decisions are promoted.

The protocol is intended to make plural forms of human knowing legible to KPGS without pretending they are interchangeable or equally authoritative for every question.

---

Why this belongs beside POP / RUNE

Existing KPGS work already distinguishes:

• POP — observe policy state and stop claimed progress exceeding evidenced progress;
• PKA — classify known / partial / unknown / conflicting state;
• POC vs FOC — test claims against reality;
• RUNE / Algiz — verify authority / endorsement boundaries;
• GSMB — preserve provenance, continuity, and receipts;
• FEP — evolve from preserved evidence rather than rewriting history.

HELP adds the missing epistemic-input layer:

REALITY
  ↓
AFFECTED HUMANS
  ↓
HELP — PLURAL KNOWLEDGE LENSES
  ↓
CONTEXTUAL MODEL
  ↓
PROPOSAL / POLICY / AGENT DECISION
  ↓
POP / PKA / RUNE / POC-FOC GOVERNANCE
  ↓
DECISION + JUSTIFICATION + RECEIPT
  ↓
LIVED-OUTCOME FEEDBACK
  ↓
GSMB / FEP EVOLUTION

---

Candidate HELP lenses

These are governance lenses / epistemic classes, not a claim that every category is a psychometrically established "intelligence" in the same scientific sense.

1. Empirical intelligence

   • measured, reproducible, observational, statistical evidence;
   • experiments, surveys, administrative data, verified outcomes.

2. Technical intelligence

   • engineering, legal, scientific, operational, economic, and domain-specialist knowledge;
   • competence must remain scoped to the domain.

3. Local / lived intelligence

   • situated knowledge from people actually affected by a system;
   • friction, informal practices, access constraints, context invisible to coarse categories.

4. Artistic intelligence

   • music, narrative, visual art, performance, storytelling and other forms that can expose affective, qualitative, symbolic, or culturally encoded realities that formal metrics may omit.

5. Spiritual / ethical intelligence

   • meaning, purpose, conscience, values, dignity, faith, existential reasoning, and claims about what a system ought to preserve or refuse;
   • must be represented without silently converting a contested philosophical/psychological construct into settled scientific fact.

---

Core invariant: relevance ≠ interchangeability

HELP MUST preserve:

• provenance;
• epistemic class;
• competence boundary;
• evidence strength;
• disagreement;
• affected-population relationship;
• decision impact;
• contestability.

Examples:

A SONG               != LEGAL PROOF
A LEGAL OPINION      != LIVED EXPERIENCE
A DATASET            != COMPLETE HUMAN ONTOLOGY
LIVED EXPERIENCE     != AUTOMATIC POPULATION ESTIMATE
SPIRITUAL CONVICTION != EMPIRICAL CAUSAL PROOF
TECHNICAL EXPERTISE  != MORAL AUTHORITY

But each may still reveal something consequential that another lens cannot.

---

Policy / legislation candidate flow

REALITY
  → affected humans
  → plural HELP lenses
  → contextual model
  → proposal
  → public participation
  → material-change detection
  → re-participation gate where legally / procedurally required
  → consideration
  → decision
  → justification
  → immutable / reconstructable receipt
  → lived-outcome feedback
  → POP observation
  → cloud update / FEP evolution

The research problem is not merely whether people were invited.

The stronger question is:

Was the decision model actually capable of being changed by relevant knowledge from outside the institution's starting ontology?

---

External prior-art baseline — do NOT claim novelty yet

Initial literature scan shows strong adjacent bodies of work that HELP must distinguish itself from before any novelty claim:

• Epistemic justice — studies how people can be wronged in their capacity as knowers; policy research has examined how evidence hierarchies can undervalue patient/professional testimony.
• Knowledge co-production / participatory governance — integrates stakeholder knowledge into actionable policy and governance; literature warns that participation can reproduce power asymmetries if plural knowledge cannot genuinely contest the process.
• Transdisciplinarity — explicitly integrates academic knowledge with extra-scientific actors and practitioners for complex societal problems.
• Indigenous Knowledge Systems (IKS) — South African law and scholarship already recognize governance, innovation, and development value in indigenous knowledge.
• Arts-based research (ABR) — recent policy scholarship argues that artistic inquiry can surface qualitative and affective realities filtered out by technocratic measurement.
• Spiritual intelligence (SI) — there is a growing but conceptually contested literature; reviews find multiple cultural definitions and continuing debate about its status, dimensions, and universality.

Representative starting references:

• Turnhout et al., The politics of co-production: participation, power, and transformation.
• Schneider et al., Negotiating space for knowledge co-production, Science and Public Policy.
• South African food-governance co-production case study, Science and Public Policy 48(2).
• Swilling, Rethinking the science–policy interface in South Africa.
• South Africa, Protection, Promotion, Development and Management of Indigenous Knowledge Act 6 of 2019.
• 2024 systematic review: Multiple perspectives of spiritual intelligence.
• 2026 scoping review / concept analysis: Spiritual Intelligence.
• 2026 Environmental Science & Policy: arts-based research and the "unmeasurable mandate".

Research requirement: HELP must identify its contribution relative to these literatures, not rename them.

---

Paper 3 candidate

Working title

HELP: Human Epistemology Layer Protocol — Governing Plural Intelligence in AI and Public Decision Systems

Alternative subtitle:

From Formal Participation to Epistemic Inclusion

Core research question

Can a governance system preserve and operationalize plural forms of human knowing across policy and AI decision pipelines without collapsing them into a single evidence hierarchy, while still maintaining provenance, competence boundaries, contestability, and fail-closed promotion rules?

Candidate hypothesis

Public and AI governance become vulnerable to epistemic exclusion when only administratively legible or technically measurable knowledge can alter the decision model, even where participation is formally present.

Secondary questions

1. What makes a knowledge contribution relevant without making it authoritative outside its competence boundary?
2. How should HELP represent conflict between empirical, technical, lived, artistic, and spiritual/ethical knowledge?
3. Can affected-person testimony alter a model without becoming unbounded anecdotal authority?
4. How should artistic inputs be preserved as evidence of meaning/experience without being promoted into factual proof they cannot support?
5. How should spiritual/ethical inputs influence value boundaries while remaining distinguishable from scientific evidence?
6. What machine-readable provenance is needed so a stateless verifier can reconstruct which knowledge changed which decision and why?
7. How does HELP interoperate with POP's policy-state observability without turning RUNE into a policy engine?
8. What is the smallest POC that can demonstrate epistemic inclusion rather than merely stakeholder presence?

---

Proposed research artifact before any code

Produce a HELP_RESEARCH_RECEIPT with:

help_input:
  source_id:
  source_actor:
  affected_population_relationship:
  epistemic_class:
    - empirical
    - technical
    - lived
    - artistic
    - spiritual_ethical
  claim:
  competence_scope:
  evidence_refs: []
  confidence:
  contestations: []
  model_change_requested:
  decision_stage:
  consideration_receipt:
  adopted_effect:
  rejected_reason:
  pka_state:
  pop_ref:
  rune_endorsement_ref:

This is a research sketch only until BGR review.

---

BGR gate

Before implementation, BGR should review at least:

• Does HELP belong inside RUNE, beside RUNE, or only at the KPGS / POP layer?
• Would a knowledge_input / epistemic_input subject type expand RUNE beyond its endorsement mandate?
• What claims, if any, should RUNE endorse: authenticity of source, consideration receipt, authority to promote, or none of these?
• How should HELP avoid converting "more perspectives" into consensus?
• What fail-closed rule applies when a material knowledge class is absent or procedurally excluded?
• What is the exact boundary between legally required public participation and KPGS's broader research notion of epistemic inclusion?
• Does implementation duplicate POP, PKA, FEP, or existing Introduction-to-MCP schemas?

Until BGR answers and owner authority promotes implementation:

HELP_CONCEPT = CANDIDATE
HELP_PAPER_3 = RESEARCH_ACTIVE
RUNE_CODE = HOLD
RUNE_SCHEMA_MUTATION = HOLD
PUBLIC_CANON = HOLD

---

Acceptance criteria for this RFC phase

• HELP is ingested into Cloud KPGS as a candidate, not canon.
• Prior-art matrix completed across epistemic justice, co-production, transdisciplinarity, IKS, arts-based research, participatory governance, and spiritual-intelligence literature.
• Definitions distinguish knowledge lens, evidence, authority, and decision right.
• At least one real policy case is mapped through HELP without claiming HELP caused or could have prevented the real-world outcome.
• Paper 3 has a bounded research question, testable propositions, method, limitations, and falsification criteria.
• BGR issues an explicit review state before any RUNE implementation.
• No public post claims HELP as proven, unique, implemented, or constitutionally required before the evidence exists.

---

Canonical internal lineage

• Introduction-to-MCP/governance/kpgs-vnext/continuity/2026-09-17_POP_POLICY_OBSERVABILITY_REALITY_CLOUD_CONVERGENCE_RECEIPT.md
• Project RUNE issue RFC: Evolve Persistency × Consistency × Context into Governed AI Execution across MMAO + MAO #3 — Persistency × Consistency × Context → governed execution
• KPGS Master Agent Orchestration Hub — HELP candidate ingestion (2026-09-19)

Core evolution:

SEE PEOPLE
  ↓
ACCEPT PLURAL WAYS OF KNOWING
  ↓
PRESERVE PROVENANCE + BOUNDARIES
  ↓
ALLOW RELEVANT KNOWLEDGE TO CHANGE THE MODEL
  ↓
GOVERN THE DECISION
  ↓
RETURN OUTCOME TO REALITY

Inclusion is not putting more people inside the existing model. Inclusion is allowing reality outside the model to change the model.