Skip to content

ADR-002: 3D Epistemic Cube (E/N/M Axes)

Status: ✅ Accepted Date: November 2025 Authors: Mycelix Core Team Supersedes: N/A (v1.0 used 1D Layered Epistemic Model) Superseded By: N/A

Context

The Problem with 1D Classification

The original Layered Epistemic Model (LEM v1.0) used a single dimension (L0-L4) to classify claims based on "epistemic strength." This approach had critical limitations:

  1. Conflated Independent Dimensions: A claim's factual verifiability (is it true?), normative authority (who agrees it's binding?), and materiality (how long does it matter?) are independent properties, but were forced into a single axis.

  2. Could Not Classify Key Cases:

  3. Constitutional principles: High authority (binding on all) but unverifiable beliefs
  4. Scientific evolution: Newton vs. Einstein both had high empirical validity but different consensus

  5. Ephemeral vs. permanent data: A "like" vs. a copyright claim have different lifespans regardless of verification method

  6. State Management Confusion: No clear way to determine which claims should be pruned from the DKG vs. preserved forever.

Why Now?

With the v6.0 architecture planning and Constitution v0.25 drafting, we needed a truth framework that could: - Support scientific evolution (superseding claims) - Enable intelligent state pruning (M-Axis) - Separate factual disputes (E-Axis) from governance disputes (N-Axis) - Classify everything from ephemeral messages to mathematical proofs

Decision

We replaced the 1D Layered Epistemic Model with a 3-dimensional "Epistemic Cube" framework.

All claims are now classified along three independent axes:

1. E-Axis (Empirical Verifiability)

Question: How do we know this claim is true? - E0: Null (unverifiable belief) - E1: Testimonial (personal attestation) - E2: Privately Verifiable (Audit Guild) - E3: Cryptographically Proven (ZKP) - E4: Publicly Reproducible (open data/code)

2. N-Axis (Normative Authority)

Question: Who agrees this claim is binding? - N0: Personal (self only) - N1: Communal (local DAO) - N2: Network (global consensus) - N3: Axiomatic (constitutional/mathematical)

3. M-Axis (Materiality)

Question: How long does this claim matter? - M0: Ephemeral (discard immediately) - M1: Temporal (prune after state change) - M2: Persistent (archive after time) - M3: Foundational (preserve forever)

Example Classifications: - A "like": (E0, N0, M0) - Unverifiable, personal, ephemeral - Passed MIP: (E0, N2, M3) - Unverifiable belief, network consensus, permanent - Mathematics: (E4, N3, M3) - Reproducible, axiomatic, permanent

Consequences

Positive Outcomes ✅

  1. Can Classify Everything: From ephemeral UI events to constitutional principles, every claim type has a clear coordinate.

  2. Enables Scientific Evolution: Claims can SUPERCEDE others (Einstein supersedes Newton) with clear graph relationships in the DKG.

  3. Intelligent State Pruning: M-Axis determines storage strategy (M0=discard, M1=prune, M2=archive, M3=preserve).

  4. Separate Dispute Paths:

  5. E-Axis disputes (fraud, counterfeit) → Member Redress Council
  6. N-Axis disputes (constitutional challenges) → Constitutional process

  7. M-Axis disputes (archival policy) → Knowledge Council

  8. Clear Governance Integration: N-Axis maps directly to governance tiers (N1=Local DAO, N2=Global DAO, N3=Constitution).

  9. Backward Compatible Schema: Upgraded Epistemic Claim Schema v2.0 adds epistemic_tier_e, epistemic_tier_n, epistemic_tier_m without breaking existing fields.

Negative Outcomes / Trade-offs ⚠️

  1. Increased Complexity: Requires understanding 3 axes instead of 1. Steeper learning curve for new contributors.

  2. Migration Effort: All existing claims must be reclassified from 1D (L0-L4) to 3D (E, N, M). Automated but requires validation.

  3. UI Challenge: Harder to visualize 3 dimensions than 1 in user interfaces. Requires thoughtful UX design.

  4. Schema Upgrade: Breaking change to Epistemic Claim Schema (v1.1 → v2.0). All clients must upgrade.

Neutral Impacts 🔄

  1. Documentation Rewrite: Extensive updates to Epistemic Charter, Architecture docs, and tutorials.

  2. Training Required: Community education needed to understand and use the 3D model.

  3. Terminology Shift: "Epistemic tier" becomes "E-Axis coordinate" + "N-Axis coordinate" + "M-Axis coordinate".

Alternatives Considered

Alternative 1: Enhanced 1D Model (L0-L9)

Description: Keep 1D approach but add more tiers to capture nuance. Pros: - Simpler conceptual model - No schema breaking changes - Easier to visualize

Cons: - Still conflates independent dimensions - Would need L0-L9 × 3 = 27+ tiers to capture all cases - Does not solve state pruning or scientific evolution - Decision: Rejected because it doesn't solve the core problem.

Alternative 2: 2D Model (Empirical + Authority)

Description: Use 2 axes (E-Axis and N-Axis) but ignore Materiality. Pros: - Simpler than 3D (2 axes instead of 3) - Solves factual vs. governance separation - Enables scientific evolution

Cons: - No solution for state pruning - DKG would grow unbounded with ephemeral data - Still missing a critical dimension - Decision: Rejected because M-Axis is essential for scalability.

Alternative 3: Do Nothing (Keep 1D Model)

Description: Continue using LEM v1.0 with workarounds for edge cases. Pros: - No migration cost - No learning curve - No schema changes

Cons: - Cannot classify constitutional principles properly - Cannot support scientific evolution - No state pruning strategy - Growing technical debt - Decision: Rejected because limitations block v6.0 features.

Implementation Notes

Migration Path:

  1. Phase 1: Schema Upgrade (Week 1)
  2. Deploy Epistemic Claim Schema v2.0
  3. Update DKG validation rules

  4. Create migration scripts for existing claims

  5. Phase 2: Charter Publication (Week 2)

  6. Publish Epistemic Charter v2.0
  7. Update all cross-references

  8. Announce to community

  9. Phase 3: Client Upgrades (Weeks 3-4)

  10. Update all official clients
  11. Provide migration tools

  12. Grace period for 3rd-party clients

  13. Phase 4: Education (Ongoing)

  14. Create tutorials and visualizations
  15. Host community workshops
  16. Update developer documentation

Technical Requirements: - DKG nodes must support 3-axis queries - Clients must validate all 3 axes - UI components for 3D visualization

Success Metrics

Quantitative: - 100% of constitutional claims properly classified as (E0, N3, M3) by Week 4 - 90% of clients upgraded to v2.0 schema by Week 8 - DKG query performance remains <100ms with 3-axis filtering - State pruning reduces storage by 40% for M0/M1 claims

Qualitative: - Community feedback: "This finally makes sense" vs. "Too complicated" - Developer onboarding time for epistemic concepts - Reduction in classification disputes

Timeline: Evaluate after 3 months (February 2026)

References

Related Documents: - Epistemic Charter v2.0 - Epistemic Charter v1.0 (superseded) - Constitution v0.24 - DKG Architecture

Research: - "Epistemic Logic and Distributed Knowledge" (Fagin et al., 1995) - "The Social Construction of Reality" (Berger & Luckmann, 1966) - "Information Architecture for the World Wide Web" (Rosenfeld & Morville, 2006)

Discussions: - Original proposal: Internal design doc (October 2025) - Community feedback: (coming in Constitution v0.25 discussion)

Revision History

Date Author Change
2025-11-10 Mycelix Core Team Initial draft
2025-11-10 Mycelix Core Team Accepted

📍 Navigation: ← ADR Index | ↑ Docs Home | Next ADR →