Emergent Constraint Fields Are Causally Active But Do Not Outperform Fixed External Rules: A Preregistered Null on Passive Emergence as a Governance Strategy

Protocol 6 of the Ethics as Emergent Constraint Response series. Tests whether constraints that emerge from agent co-constitution produce better alignment-relevant behavioral outcomes than externally imposed fixed rules. Four-condition preregistered design (emergent local, emergent global, fixed external, unconstrained). 200 confirmatory runs (50 seeds × 4 conditions, 500 epochs each). Key findings: mechanistic prediction strongly confirmed (median entropy-SSS correlation r = −0.680, p < 0.001); primary behavioral claim not confirmed (A vs. C, p = 0.069); H3 reversed (global field perception produced more behavioral variance than local, contrary to all committee predictions); H4 not confirmed. Governance conclusion: emergent constraint fields are causally active but do not outperform well-calibrated fixed external rules. Passive emergence is insufficient as a governance strategy. Preregistration DOI: 10.5281/zenodo.19297509. Repository: github.com/btisler-DS/emergent-constraint-marl. This version adds supplementary confirmatory artifacts, including the analysis JSON, findings report, prediction scoring report, locked configuration, and analysis script. It also adds the related Protocol 3 preregistration and results records to the identifier graph. The scientific claims and main paper PDF are unchanged from the original P6 results publication.