Prospective protocol details a reduced three-criterion testing framework for causal emergence benchmarks, aiming to restrict false positive rates below previous failure thresholds.
This record contains the frozen prospective protocol for version 0.3.0 of the Operational Transition Test (OTT) Experimental Crucible. Context. The v0.2 frozen protocol (DOI 10.5281/zenodo.22116064) was publicly timestamped before execution; its prospective run was then published unmodified (DOI 10.5281/zenodo.22116481) and the primary frozen hypothesis H1 failed: the context_fragile_macro negative family reached FPR = 0.3750 against the frozen per-family limit of 0.10. v0.3 is the prospective reduction response to that specific falsifier; the v0.2 result record remains unchanged. What v0.3 changes. The six-criterion v0.2 battery (F, K, E, D, R, G) is reduced to a three-criterion V–C–X core: V — scope-wide interventional validity; C — robust control-cost compression; X — attributable constrained operational expansion. The reduction decisions are documented inside the archive (protocol/core_definition.json, docs/ALGEBRAIC_AUDIT_v0.2.md). Freeze statement. All v0.3 thresholds, benchmark families, seed derivation rules, development/evaluation allocation, resource accounting, baselines, ablations, and scoring rules are frozen before prospective execution. NO v0.3 prospective development/evaluation run has occurred. The archive contains run-lock machinery (template only); no RUN_AUTHORIZATION.json exists in this deposit. Non-claims. VCX is not a new theory of causal abstraction; threshold values are benchmark decision thresholds, not universal constants; passing the prospective benchmark would establish discriminative utility in the benchmark, not a universal law of emergence. Integrity (SHA-256). ZIP archive: 28a7d5cef7e3b2791781b0acec0fa495e9a02caeeb9f213b1b12153467bb9c36; manuscript PDF: 45119ef10bb48547059cd8a54e439a15dc78ef12147069e1c53454f10ca7cd3c; canonical protocol root: 44cc90e34397b83a4bce29ab9307511dcdf78e5a22b97bc39f03ecd4bfa29b06 (all 32 protocol-manifest files independently verified before publication; uploaded files confirmed byte-identical). Licensing. Record license CC BY 4.0 (scientific text, protocol, documentation, benchmark data). Source code (src/, tests) remains under Apache License 2.0 per LICENSES.md inside the archive.
No takes yet. Share an insight, caveat, or question.
Sławomir Grzegorz Gątkowski (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: