Skip to content

Validation

The project uses a validity argument, not a single validation stamp. Start with publication status and evidence and the validation doctrine.

Structural validation

python scripts/validate_obelisk.py

Checks stable IDs, references, prerequisites/cycles, branch invariants, life-pillar structure, formation/AI modes, source links, and preservation of the 417-row All Souls corpus.

Candidate release / claim maturity

python scripts/check_obelisk_release_readiness.py
python scripts/check_obelisk_release_readiness.py --claim empirical
python scripts/check_obelisk_release_readiness.py --claim longitudinal

The first command checks recorded prerequisites for a candidate release within its programmed scope. It is not an editorial review, bibliographic audit or proof of learning. The empirical and longitudinal checks fail while the required evidence is absent.

Evidence layers

  • construct/content coverage
  • external human/professional proxy evidence
  • response-process evidence
  • unseen transfer
  • judgment and adaptive-agency performance (including adversarial/surprise stress cases)
  • longitudinal real-life evidence

All Souls questions offer material for a proposed intellectual-judgment proxy, not evidence that this curriculum produces that judgment. The imported mappings have unverified independent hold-out provenance and do not establish institutional endorsement or whole-life competence. See the hold-out research program and proposed general examination.