Back to PJC Framework

Public technical method · Research Preview v0.6

PJC Judgment Record

Specification: 0.6
Rule: commit before the outcome, never overwrite the original, and settle from a pre-specified source.

1. Standard fields

FieldRequiredMeaning
Record IDYesGlobally unique identifier
Created AtYesTime and time zone
SubjectYesPerson accountable for the judgment
ContextYesDecision setting and scope
PropositionYesClaim a later outcome can test
ConfidenceYesSubjective probability from 0–100%
RationaleYesPrincipal supporting reasons
AlternativeYesAt least one plausible alternative
Observation WindowYesStart and end of observation
Outcome SourceYesPre-specified data or adjudication source
Falsification ConditionYesResult that would contradict the judgment
Stop ConditionOptionalCondition requiring a halt or escalation
Committed AtYesTime the original record was frozen
Actual OutcomeAt settlementObserved result
Settlement StatusAt settlementSupported, contradicted, inconclusive, or invalidated
Settlement EvidenceAt settlementLink, extract, or signed decision
AdjustmentAt settlementConcrete change for the next cycle

2. Lifecycle

StateEditable scope
DraftAll fields
CommittedAppendices only
Awaiting OutcomeNew evidence and environmental changes only
SettledReview comments only
InconclusiveExcluded from accuracy calculations
InvalidatedRetained but excluded from calibration

Every state transition keeps actor, time, and reason.

3. Blank template

record_id:
created_at:
subject:
context:
proposition:
confidence:
rationale:
alternative:
observation_window:
outcome_source:
falsification_condition:
stop_condition:
committed_at:
appendices: []
actual_outcome:
settlement_status:
settlement_evidence:
adjustment:

4. Worked example

At commitment

FieldValue
Record IDPJC-2026-001
PropositionWithin 30 days, the new onboarding flow will raise first-canvas success to at least 45%
Confidence70%
RationaleFour of five recent interviews showed hesitation at the blank-canvas entry
AlternativeDrop-off is mainly caused by installation and performance
Observation Window1–30 September 2026
Outcome SourceProduction create_canvas_success event
Falsification ConditionAt least 500 eligible users and success below 45%
Stop ConditionRoll back if crash rate rises by one percentage point

At settlement

FieldValue
Actual Outcome612 eligible users; 42.8% success
Settlement StatusContradicted
Settlement EvidenceFrozen export dated 1 October 2026
AdjustmentSeparate installation failure from blank-canvas hesitation next cycle

A contradicted record is useful: a concrete proposition held at 70% confidence missed its threshold.

5. Integrity check

Before commitment, verify that the proposition has an object and time, confidence is probabilistic, the outcome source is fixed, falsification mirrors the claim, the alternative is genuine, and the human subject actively confirms submission.

Before settlement, verify that the original source was used, original fields remain intact, inconclusiveness is recorded honestly, and environmental changes were not silently folded into the original claim.