Prospective recording
Fix the judgment, basis, confidence, observation window, and falsification conditions before the outcome is known.
Produces an append-only commitment record.AN OPEN SEDES MENTIS TECHNICAL METHOD · RESEARCH PREVIEW v0.6
PJC FRAMEWORK
Prospective Judgment Calibration Framework
PJC Framework
Before the outcome is known, write down what you expect, why you expect it, and what would show you were wrong. When the outcome arrives, review it against the original criteria.
PJC is an open technical method for low-validity human–AI agent environments: tasks where feedback is delayed and the relationship between cues and outcomes is weak or unstable. It organises this process through prospective recording, outcome settlement, and temporal calibration.
01 · PROBLEM
Task completion does not mean human judgment improved
Fluent output can manufacture certainty, and people can rewrite their earlier judgment once an outcome is known. PJC preserves the judgment and supporting reasons written before the outcome so later reality can test them.
02 · METHOD
Three linked mechanisms
Outcome settlement
After the observation window, settle against a pre-specified external outcome source.
Distinguishes supported, contradicted, inconclusive, and invalid records.Temporal calibration
Compare confidence, judgment types, and outcomes across repeated cycles.
Provides material for examining recurring judgment biases.03 · BOUNDARIES
Scope of use
Use when
- Outcomes arrive later
- Judgments repeat
- Outcome sources can be pre-specified
- Confidence needs calibration
Do not use as
- Pure creative expression
- Values without observable outcomes
- One-off incomparable choices
- Using records as performance scoring
04 · EVIDENCE
Current evidence status
The method structure exists, and internal operability is under evaluation; calibration effects and business benefits have yet to be validated. PJC currently makes no promise of higher efficiency, accuracy, or team performance.
- Method structure
- Specified
- Internal operability
- Under evaluation
- Cross-context validity
- No empirical test started
- Calibration effect
- No empirical test started
- External peer review
- Not conducted
05 · DOCUMENTATION
Method documents
01
Practitioner guide
Run a complete PJC cycle.
02Judgment record
Fields, states, and a worked example.
03Research note
Sources, mechanisms, and testable hypotheses.
04Workflow integration
Structural constraints for agent systems.
05Evidence and cases
Current status and pilot standards.
06Claim register
Evidence level and falsification criteria.
07Changelog
Version and terminology history.
08License
CC BY 4.0 open license.