Conditional predictions and qualified experimental comparison

Assume the Axiom of Countable Choice (${\mathrm{A}\mathrm{C}}_{\omega }$, The Axiom of Countable Choice (${\mathrm{A}\mathrm{C}}_{\omega }$)) exactly as inherited from clause (a) of The integral form of the vacuum Maxwell equations and its equivalence with the smooth differential form, which is used in clause (a)(i) below with its piecewise $C^1$ boundary and surface-measure conventions; no other choice principle is used. Let the laboratory region, its inertial chart, the fields $E,B$ and the source fields $\rho ,J$ be those named in The integral form of the vacuum Maxwell equations and its equivalence with the smooth differential form, with its constants and units, and let $R$ be a record of Report-backed operational evidence with uncertainty of dimension $d=1$ (primary source, apparatus-and-preparation description, recorded data or reported observations, calibration model, uncertainty record, analysis-assumption list, interpretation). Write $I_t$ for the laboratory time interval. Let $I_{\theta }\subseteq \mathbb{R}$ be a nonempty open parameter interval, ${\theta }_0\in I_{\theta }$, and $f:I_{\theta }\to \mathbb{R}$ a declared comparison map of class $C^1$ such that $f({\theta }_0)=0$ and $f^{\mathrm{\prime }}(\theta )\ge m>0$ for every $\theta \in I_{\theta }$. Suppose the record's declared apparatus-and-calibration dictionary identifies the comparison quantity with the deviation $y=f(\theta )$ of a declared model family from the exact framework reference prediction $y_0$ of clause (a), and declares the recorded deviation $\widehat{y}$ and the bounds $\mathrm{\mid }\widehat{y}\mathrm{\mid }\le s$ and $\mathrm{\mid }y-\widehat{y}\mathrm{\mid }\le \delta$ with $s,\delta \ge 0$.

(a) [Exact framework reference predictions] (i) [Static branch] Assume the smooth branch of the adopted vacuum Maxwell equations with the regularity hypotheses of The integral form of the vacuum Maxwell equations and its equivalence with the smooth differential form (fields jointly $C^1$ on $\mathrm{\Omega }\times I_t$ with the additional derivatives appearing in the displayed equations existing and continuous there; for instance $E,B$ jointly of class $C^2$ and $\rho ,J$ of class $C^1$). Let $(V,{\nu }_V)$ be a volume test geometry of Fixed volume and patch test geometries, outward normals and induced orientations and let $t\in I_t$ be such that $\rho (\cdot ,t)$ vanishes identically on $V$. Then clause (a) of The integral form of the vacuum Maxwell equations and its equivalence with the smooth differential form gives ${\oint }_{\mathrm{\partial }V}E(\cdot ,t)\cdot {\nu }_{V}dS=\frac{1}{{\epsilon }_0}{\int }_V\rho (\cdot ,t)d{\lambda }_3=0.$ On the smooth branch every admitted boundary $\mathrm{\partial }V$ whose enclosed volume $V$ is charge-free has vanishing electric flux; charge-freeness of the surface alone is insufficient, and the exact framework prediction for the enclosed-flux observable is the value $y_0=0$.

(ii) [Wave branch] Let $(E,B)$ be the plane-wave solution of Period-averaged energy density, Poynting vector and momentum flux of a plane wave with data $(k,\omega ,e)$ there, and let $⟨S⟩$ be the period average of its Poynting vector. Then that item gives the signed identity $⟨S⟩\cdot n=\frac{{\epsilon }_{0}c}{2}\mathrm{\mid }e{\mathrm{\mid }}^2(\widehat{k}\cdot n)$ for every unit vector $n$; if the declared dictionary fixes the comparison quantity as the magnitude $\mathrm{\mid }⟨S⟩\cdot n\mathrm{\mid }$ of the normal period-averaged flux (a declared intensity convention), then the exact framework prediction is $y_0=\frac{{\epsilon }_{0}c}{2}\mathrm{\mid }e{\mathrm{\mid }}^2\mathrm{\mid }\widehat{k}\cdot n\mathrm{\mid }\mathrm{.}$ The absolute value is part of that declared magnitude convention and is not a consequence of the signed identity.

(b) [Conditional confinement] Under the stated identifications, every parameter value $\theta \in I_{\theta }$ compatible with the declared model family, the apparatus-and-calibration dictionary and the two declared bounds satisfies $\mathrm{\mid }\theta -{\theta }_0\mathrm{\mid }\le \frac{s+\delta }{m}\mathrm{.}$ Equivalently, a proposed $\theta$ violating this bound is incompatible with the conjunction (specified family, dictionary, calibration bound, recorded bound); the conclusion is a deterministic statement about the declared bounds.

(c) [Status and qualifications] The conclusion is conditional on the declared family, dictionary, calibration and recorded bounds and on the declared monotonicity margin $m$; it is not a sampling law, a likelihood, a confidence level, a coverage probability, a $p$-value or a probability that any model or framework is true, and it asserts no rejection error rate. Agreement of a record with a specified family, like disagreement with it, concerns the conjunction of that family with the apparatus, calibration and analysis assumptions; a discrepancy may lie in the family, the dictionary, the calibration or the analysis, and neither agreement nor disagreement proves or deductively falsifies the adopted framework. The three primary reports recorded on this page supply, in their recorded read extents, no calibrated response transfer function, no trial-level dataset and no declared sampling law; accordingly this item asserts no numerical bound and performs no numerical application for any of them, and their own uncertainty records carry their source limitations. Empirical support and its constraints are nonlogical relations, not dependencies, and any consumer of this item inherits $s$, $\delta$, $m$, the declared family, the record's conditions and the source limitations.