Beyond 1,419,857 Paths: From Wen et al.'s Experimental Test of Feynman Path Integrals to the Measurement-Architecture Boundary

Yong-Li Wen et al. (2026) experimentally test Feynman’s two path-integral postulates with single photons by measuring complex propagators and reconstructing 17^5 = 1,419,857 discrete path skeletons. This record accepts that experimental result within its declared paraxial domain and asks a different question: when a measurement architecture maps a richer source space into a reduced readout, which distinctions are actually measured, which are generated by deterministic composition, and which remain outside the jurisdiction of the declared task? ORIGAS Cabin 1/2 Device IX v9.5 formalizes this question as a measurement-architecture boundary. It develops task-relative representational jurisdiction, fibre-based sufficiency criteria, closure under intervention, covariance-aware model-manifold tests, frozen provenance, and independent null/recovery protocols. The public Wen/Dryad dataset is used as the empirical baseline; reconstruction depth is explicitly distinguished from independent source width and from representational completeness. The record freezes prospective tests rather than retrospectively fitting them: the P9-C/Q5/Q7 photon closure programme, a Point-0 balance prediction centered on an operationally definedμ = 1/2, +t/−t, K × T OFF/ON/recovery law, and the previously frozen Double RingCross outer-Solar-System prediction as a chronologically independent external adjudicator. A separate author-frozen layer records sociotechnical and AI-jurisdiction predictions, including the France 2026–2027 case study. These are explicitly separated from the empirical Wen photon results.** No Device IX calculation is presented as a measured validation of a physical World B, Complete EPR, \(+t/-t\), \(K\), \(T\), or the Double RingCross. Bell/Aspect, Connes and Wen are treated as valid within their declared tasks. The narrower claim is that success of a reduced representation for one task does not establish injectivity, sufficiency, or closure for every enlarged task. Device IX is therefore experimentally adjudicable, but its prospective physical branches remain open until independently tested. The central methodological rule is: a reduction may be valid without being complete; the relevant question is whether the representation preserves the distinctions required by the task.Direct experimental test of Feynman’s path integral postulates with single photons | Science Advances