ECR-000001 Interpretation Rules

• Do not treat calibration as validation.
• Do not strengthen broader Framework Engineering hypotheses on calibration evidence alone.
• Record observations separately from interpretations.
• Product relevance observations are not product validation.
• Recognition must be treated as a threat to validity.
• Primitive agreement should be compared against structural and constraint agreement.
• If one representation fails but another remains stable, split the hypothesis rather than forcing a single conclusion.
• If Bloom appears highly procedural, investigate instrument bias first.
• If a packet produces rich output for a static artifact, investigate instrument bias before claiming discovery.
• If schema sections are ignored, treat that as an instrument problem before treating it as a theory result.
• Respect pass/fail criteria before considering the instrument ready.