TriadicFrameworks
Panoramica

SARG_Error_Taxonomy.md

SARG Error Taxonomy
A unified triadic classification system for structural, anchor, lineage, and novelty‑driven errors across any SARG‑described substrate.

1. Overview#

The SARG Error Taxonomy provides a universal grammar for describing, classifying, and resolving errors that arise during:

  • substrate parsing
  • lens application
  • invariant extraction
  • resonance mapping
  • lineage traversal
  • cross‑substrate synthesis

Errors are grouped into four primary classes, each aligned with a SARG operator:

Class Operator Meaning
S‑Errors Structure Something is malformed, missing, or contradictory in the substrate or its declared structure.
A‑Errors Anchors Resonance anchors, axes, or cluster mappings fail or conflict.
L‑Errors Lineage Lineage chains, inheritance paths, or cross‑substrate references break.
H‑Errors High‑Novelty The system encounters something outside its known invariants or resonance space.

Each class contains three subtypes, forming a triadic 3×3 grid.

2. The SARG 3×3 Error Grid#

S‑Errors (Structural)#

Structural errors arise when the substrate or its declared shape cannot be trusted.

Code Name Description
S1 Missing Structure Required fields, elements, or properties are absent.
S2 Contradictory Structure Declared structure conflicts with observed structure.
S3 Unstable Structure Structure is present but inconsistent across passes.

A‑Errors (Anchor)#

Anchor errors occur when resonance anchors cannot be established or maintained.

Code Name Description
A1 Missing Anchor Expected anchor or cluster not found.
A2 Conflicting Anchor Multiple anchors claim the same element.
A3 Drifted Anchor Anchor exists but no longer matches invariants.

L‑Errors (Lineage)#

Lineage errors arise when the system cannot trace or validate lineage relationships.

Code Name Description
L1 Missing Lineage Required lineage reference not found.
L2 Broken Lineage Lineage chain exists but cannot be resolved.
L3 Cyclic Lineage Lineage forms a loop or illegal recursion.

H‑Errors (High‑Novelty)#

High‑novelty errors indicate the system has encountered something outside its known resonance space.

Code Name Description
H1 Novel Element Element not in substrate or known invariants.
H2 Novel Pattern Pattern does not match any known invariant class.
H3 Novel Resonance Resonance signature cannot be mapped.

3. Error Event Shape (Canonical)#

Every SARG error is represented using the standard event shape:

{
  "pattern_id": "string",
  "source_node": "string",
  "timestamp": "ISO-8601",
  "sarg_stage": "substrate | lens | invariants | resonance | lineage",
  "error_code": "S1 | S2 | ... | H3",
  "details": {
    "message": "Human-readable description",
    "context": {},
    "suggested_rectification": "optional"
  }
}

This shape is used by:

  • local node validators
  • the Cloud Rectification Cluster
  • lineage walkers
  • resonance mappers
  • substrate loaders

4. Rectification Actions (Summary Table)#

This table matches the one we generated earlier, now embedded in the taxonomy.

Error Code Rectification Action
S1 Request missing structure or regenerate substrate.
S2 Reconcile contradictions or choose canonical structure.
S3 Stabilize structure via re‑sampling or majority vote.
A1 Recompute anchors from invariants.
A2 Resolve anchor conflicts using priority rules.
A3 Re‑align anchor to updated invariants.
L1 Request missing lineage or fallback to root.
L2 Rebuild lineage chain using nearest valid nodes.
L3 Break cycle and re‑establish legal lineage.
H1 Add element to novelty buffer for human review.
H2 Attempt invariant generalization.
H3 Create provisional resonance signature.

5. Error Flow (SARG‑Aligned)#

[Substrate] → S‑Errors
      ↓
[Lens Application] → A‑Errors
      ↓
[Invariant Extraction] → H‑Errors (pattern-level)
      ↓
[Resonance Mapping] → A‑Errors (anchor-level)
      ↓
[Lineage Traversal] → L‑Errors

This flow mirrors the Capture.md pipeline you have open.

6. High‑Novelty Handling#

High‑novelty events (H‑class) are not “failures” — they are expansion points.

The system:

  1. Captures the novel element/pattern/resonance
  2. Stores it in the novelty buffer
  3. Attempts generalization
  4. Emits a provisional signature
  5. Awaits human confirmation

This is how SARG evolves.

7. Example Error Events#

Example 1 — Missing Structure (S1)#

{
  "pattern_id": "latin_A",
  "source_node": "substrate_loader",
  "timestamp": "2026-04-20T14:00:00",
  "sarg_stage": "substrate",
  "error_code": "S1",
  "details": {
    "message": "Substrate missing required field: elements",
    "context": {}
  }
}

Example 2 — Conflicting Anchor (A2)#

{
  "pattern_id": "letter_O",
  "source_node": "resonance_mapper",
  "timestamp": "2026-04-20T14:01:00",
  "sarg_stage": "resonance",
  "error_code": "A2",
  "details": {
    "message": "Element O claimed by both curved_forms and rotational_symmetry clusters.",
    "context": {}
  }
}

8. Notes for Implementers#

  • Keep error messages short, structural, and substrate‑agnostic.
  • Never embed domain‑specific assumptions in error codes.
  • Use triadic grouping to maintain SARG coherence.
  • Emit H‑class errors liberally — they are how the system learns.
  • Maintain lineage integrity; L‑errors are the most expensive to repair.

If you want, AI can also generate:

  • /docs/SARG/error/SARG_Error_Examples.json
  • /docs/SARG/error/rectification_flow.md (diagram‑first)
  • /docs/SARG/error/error_signatures.json (machine‑readable)
  • or fold all of this into a SARG Error Pack for the repo.

Updated

Powered by Docsbook