This record contains version 0.5 of the Epsilon–Theta (ET) Framework Meta-Specification: a framework-level specification for baseline-relative deviation analysis in finite bipartite Bell scenarios. The ET framework defines a residual-first language for comparing observed Bell data with an explicit baseline probability model. Its primary object is the full observable residual table, from which scalar radii, compressed static summaries, Bell-functional projections, no-signalling diagnostics, and temporal diagnostics may be derived under declared conventions. The framework separates two complementary layers of analysis. The epsilon layer describes static, baseline-relative deviations in conditional probability tables. The theta layer provides a generic contract for temporal, order-dependent, or higher-order diagnostics, such as memory, drift, block dependence, and settings-time confounding, under declared calibration protocols. This Meta-Spec defines the framework-level contract: scope, baseline modes, residual-first principle, conformance classes, reporting requirements, calibration discipline, extension rules, and the relationship between the general framework and named scenario packs. It does not define a universal compressed residual vector, universal Bell functional, or universal temporal statistic. Those objects are intentionally delegated to scenario-specific packs. The document is not a new physical theory, not a new Bell inequality, and not a replacement for Bell-valid significance analyses, loophole-specific experimental audits, device-independent security proofs, or quantum-set compatibility tools. Instead, ET is intended as a complementary reporting, diagnostic, and sensitivity framework for declared classes of baseline-relative observable departures. The canonical 2x2x2 binary-outcome scenario is specified separately in the ET-CS-222 scenario pack and its companion supplements. Future versions of the Meta-Spec may extend the framework-level contract to additional Bell-scenario families, such as multipartite, network, or adaptive settings.
Azat Ahmedov (Sun,) studied this question.