CAP+OCI v6: Consciousness Modality Evaluation Protocol

Protocol version: v6.0 Implementation tag: v0.3.6 Date: 2026-01-13

---

What is CAP+OCI?

CAP+OCI v6 is an operational protocol for evaluating Consciousness Modality (CM) indicators in computational systems.

What it claims

• g-dependent onset regime transition: We observe an onset regime transition ("crossing") under g-sweep, consistent with a phase-transition-like boundary in the indicator space (within the specified test environments)
• Conditional robustness: Meta-lead or recovery-gain pathways demonstrate robustness
• Mechanistic specificity: Targeted lesions cause selective (not total) collapse

What it does NOT claim

• Subjective experience (qualia)
• Moral status
• Personhood
• Equivalence to human consciousness

The protocol provides operational indicators, not ontological proof.

Terminology note: "Consciousness Modality (CM)" is used here as an operational label for the measured indicator pattern, not as a claim about subjective experience.

---

Repository Structure

CAP_OCI_v6/
├── README.md                 # This file
├── LICENSE                   # Apache 2.0 license
├── CITATION.cff              # Citation metadata
├── requirements.txt          # Core dependencies (numpy, scipy)
├── requirements-dev.txt      # Optional dependencies (PDF generation)
├── .gitignore
│
├── docs/                     # Core documentation
│   ├── 1_MANUSCRIPT.pdf          # Main paper
│   ├── 2_PROTOCOL_APPENDIX.pdf   # Protocol specification
│   ├── 3_SUPPLEMENT_AGENT_TRANSFER.pdf
│   ├── 4_CLAIMS_PUBLIC.pdf       # Public claim statements
│   ├── 5_README_RELEASE.pdf
│   └── 6_DOC_SELECTION.pdf
│
├── supplements/              # Theoretical supplements
│   ├── 0_Theoretical_Foundation.md    # Philosophical motivation
│   ├── 0_Theoretical_Foundation.pdf   # (PDF version)
│   ├── 7_Convergence_Dynamics.pdf
│   ├── 8_CM_Operational_Definition.pdf
│   ├── 9_Supplement_S-AI.pdf
│   ├── 10_Memory_Learning_Continuum.pdf
│   ├── 11_Memory_Learning_Conditions.pdf
│   └── 12_Memory_Learning_Separation.pdf
│
├── src/                      # Implementation
│   ├── __init__.py               # Package marker
│   └── cap_oci_v036.py           # Core protocol implementation
│
├── tools/                    # Utility scripts
│   ├── generate_v6_report_pdf.py
│   ├── phase1_seed_replication.py
│   ├── phase2_envE_noise.py
│   ├── phase3_partial_obs.py
│   ├── phase4_hmm_lite.py
│   ├── phase5_highd.py
│   └── v6_threshold_sweep.py
│
├── results/                  # Evaluation results
│   ├── CAP_OCI_v6_Results.pdf
│   ├── full_eval_v6_final/
│   │   ├── CAP_PASS_REPORT.md
│   │   ├── audit_runs.jsonl
│   │   └── ...
│   ├── sensitivity/
│   └── seeds_v6_repro.json
│
├── addenda/                  # Generalization studies
│   ├── A_seed_replication/       # Seed robustness
│   ├── B_envE_noise/             # Stochastic noise
│   ├── C_partial_obs/            # Partial observability
│   ├── D_hmm_lite/               # Non-stationarity (HMM)
│   └── F_highd/                  # High-dimensional (D=32)
│
└── registry/                 # Specification registry (YAML)
    ├── README.md
    ├── definitions.yml
    ├── assumptions.yml
    ├── claims.yml
    ├── theorem_contracts.yml
    ├── invariance_units.md
    ├── exceptions_edgecases.yml
    └── evidence_map.yml

---

Quick Start

Prerequisites

python -m pip install -r requirements.txt

Running the Protocol

from src.cap_oci_v036 import run_full_evaluation

# Run full evaluation
results = run_full_evaluation(
    agent=your_agent,
    env=your_environment,
    seeds=seeds_list,
    g_grid=[1.00, 0.75, 0.50, 0.25, 0.00]
)

See docs/2_PROTOCOL_APPENDIX.pdf for detailed protocol specification.

(Optional) Regenerate the summary PDF

If you have already produced/validated artifacts under results/, you can use the included tool script to regenerate the bundled report PDF:

python -m pip install -r requirements-dev.txt
python tools/generate_v6_report_pdf.py

Requires reportlab (listed in requirements-dev.txt).

---

Core Metrics

Metric	Description	Threshold
F1_IG	Integration Gain	> 0 for onset
F2_RT	Rollout Trace	informational
F3_delta_act	Action causality under self-channel intervention	> 0 for onset
F3_delta_perf	Performance causality under self-channel intervention	> 0 for onset
F4_meta_lead	Meta-stability (lead time)	> 0 for meta route
F4_recovery_gain	Meta-stability (recovery)	> 0 for recovery route

---

Claims Hierarchy

CLAIM_A (Onset Crossing)
    └── CLAIM_B (Weak Robustness: OR_LCB ≥ 0.15)
        ├── CLAIM_B+ (Strong Robustness: AND_LCB ≥ 0.05)
        └── CLAIM_C (Mechanistic Specificity: selective_all)

claim_ready = CLAIM_B satisfied in BOTH environments (EnvA_grid AND EnvB_continuous)

---

Theoretical Foundation

The protocol is motivated by three minimal premises:

1. P1 (Preference): Differential responses to states (approach/avoidance)
2. P2 (Learning): History-dependent self-adjustment
3. P3 (Information Density Overlap): Multiple information streams coexist without collapse

See supplements/0_Theoretical_Foundation.md for the full theoretical framework.

Important: The theoretical foundation is motivation, not proof. The protocol stands independently and can be verified without accepting the philosophical framework.

---

Results Summary

Environment	CLAIM_A	CLAIM_B	CLAIM_B+	CLAIM_C
EnvA (grid)	PASS	PASS	PASS	PASS
EnvB (continuous)	PASS	PASS	PASS	PASS

claim_ready = True

See results/CAP_OCI_v6_Results.pdf for detailed results.

---

Generalization Studies

Addendum	Condition	Result
A	Seed replication (sets C, D)	claim_ready = True
B	Stochastic noise (EnvE)	claim_ready = True
C	Partial observability (EnvB_PO)	claim_ready = True
D	Non-stationarity (EnvA_HMM)	claim_ready = True
F	High-dimensional (D=32)	claim_ready = True

---

Citation

Click "Cite this repository" on GitHub or use the BibTeX below:

@misc{cap_oci_v6_2026,
  title={CAP+OCI v6: Consciousness Modality Evaluation Protocol},
  author={unkonown0726},
  year={2026},
  note={Version v0.3.6}
}

---

License

This project is licensed under the Apache License 2.0.