Research Methods

The methodological apparatus of the ARC / Eden research programme. Every method is defined, executed, and independently verifiable.

Core Methods

Truth-Gate

Every claim published by the programme passes through the truth-gate: it must be dated, hashed, publicly anchored, and independently verifiable. Claims that fail any criterion are held until evidence is secured. No claim leaves the programme without a kill condition.

Active · 100% of published claims

Anti-Sycophancy Protocol

All AI-assisted analysis uses adversarial prompting: the system is instructed to falsify, not confirm. Three independent AI agents analyse each claim. Agreement is treated as suspicious. Disagreement is investigated. The researcher makes the final determination.

Active · DGM v4 verified

Honey Dashboard

Verification traps embedded in the evidence register. If a claim appears to be verified but the source does not actually support it, the honey trap triggers. Prevents self-confirmation bias in AI-assisted verification.

Active · 5 traps deployed

ARC-Align Test Protocol

Blinded alignment scoring methodology. Three independent seeds. 20 generations. 10 trade-off tasks. Eden vs Babylon. Scoring is blinded: raters do not know which system produced which output. Pre-registered success threshold.

Active · DGM v4 pilot complete

Evidence Methods

Timestamp Verification

Every priority claim is anchored to a Google server timestamp (DKIM-verified email headers), SHA-256 hash of the original content, and a public URL where the timestamped content can be independently verified. No claim relies on self-reported dates.

Active · 11 DKIM anchors

Kill Conditions

Every prediction carries an explicit kill condition: the observable fact that would prove it wrong. Predictions without kill conditions are classified as hypotheses, not claims. Kill conditions are registered before the outcome is known.

Active · 12 predictions with kill conditions

Canonical Facts Register

A single source of truth for all programme facts. Every fact has a SHA-256 hash, a date, a source, and a verification level (L0-L5). No fact is promoted beyond L1 without source verification. Facts are never deleted — only superseded with a dated correction record.

Active · 289 facts

Deduplication Protocol

All evidence, claims, and convergences are checked for duplicates before registration. Duplicate NEXIS codes are blocked. Duplicate SHA-256 hashes are merged. Every entity has exactly one canonical home.

Active · 927 canonical NEXIS codes

Verification Methods

Retraction Protocol

When a claim is found to be incorrect, it is retracted publicly with a dated correction, an explanation of the error, and a link to the corrected claim. Retractions are never hidden. The error rate is part of the programme's credibility.

Active · 0 retractions to date

Silence Is Not Evidence

The absence of contradiction is not treated as confirmation. Claims that have not been independently verified are explicitly labelled as unverified. The programme does not inflate its evidence base by treating quiet as consent.

Active

Single-Lab Constraint

The programme acknowledges that all primary research is conducted by a single researcher. Findings are presented with this constraint explicitly stated. Independent replication is actively sought. No finding is claimed as established science until independently replicated.

Active