Spatial Displacement Theory · Section Paper VIII · Prepared for Submission
The Falsification Programme: Ninety-Seven Numbered Specifications, E13–E109
Melbourne, Australia · July 2026 · Atlas Chapter Seven — Experiments
Abstract. The framework carries 97 numbered experiment specifications (E13–E109) across the nuclear, cosmological, orbital, and electromagnetic sectors. Each specification states three things: the SDT prediction, the rival prediction, and the observable that discriminates. This section paper formalises the specification format, the difficulty ladder (already testable · moderate · frontier · extreme), and the programme's operating rules: predictions are pre-registered before instrumentation; instruments are validated on known answers before use; comparisons are whole-range and single-pass; verdicts — including failures — are retained on display. Worked entries are given for the convergent anchors (perihelion closure, limb deflection), the divergent discriminants (mass-independent orbital drift, substrate acoustics between shielded plates, γ-ray dispersion, the neutrino moment, post-merger echoes), and the registered kill conditions the framework exposes itself to (monopole detection, CMB spectral distortion above 50 ppm, the solar-limb fold scan).
Keywords: specification format · discriminants · pre-registration · kill conditions · difficulty ladder
1. The specification format
A specification is admissible only if the discriminant is stated as a measurement that can return either verdict. Entries whose discriminant turns out to be shared form — both frameworks predicting the same number for reasons of their own — are re-labelled convergent anchors and removed from the discriminating set (the fragment-energy fine-tune of the beam-peel campaign was retired exactly this way). The programme's verdicts are retained: a failed construction remains on display with its verdict attached, because the record is the product.
2. Convergent anchors (already testable, both frameworks land)
| ID | Observable | Record |
|---|---|---|
| E24 | anomalous perihelion precession | 42.98″/cy from ϟ = R/k² with G and M never written; measured 42.98. Convergent with the rival — independent origins, one number; the same closure recovers c to +0.0009% |
| E47 | solar-limb deflection | 1.750″ from the accumulated depth gradient vs 1.751″ measured — flat substrate, no metric curvature |
| E07 | primordial ⁷Li | deficit from α+d+t construction-grammar constraints on assembly channels; no stellar depletion invoked |
Anchors establish that the formalism reaches the precision data with less machinery; they discriminate nothing by themselves and are labelled so. The discriminating load falls on the divergent set.
3. Divergent discriminants (the framework's exposure)
| ID | SDT prediction | Rival prediction |
|---|---|---|
| E21 | mass-independent lateral drift > 10 km in free orbital descent (pressure-gradient mechanics) | geodesic motion: no such drift |
| E22 | mechanical signal propagation through evacuated space between shielded plates (substrate acoustics) | field-theoretic vacuum: null |
| E100 | frequency-dependent γ-ray group velocity from ℓP granularity (arrival-time spread vs energy) | continuum: no intrinsic dispersion |
| ν-moment | μν ≡ 0 (open winding, no ℓ ≥ 1 wake) | minimally-extended: non-zero |
| GW echoes | post-merger echoes at Δt ≈ 58 ms; no breathing mode | no echoes |
| TD03 fingerprint | rank-4 lattice observable in [1/9, 7/27] | continuum: exactly 1/5 |
| E97 | displacement-pressure inner-core seismic predictions on the existing travel-time catalogue | — |
4. Registered kill conditions
The framework names, in advance, the observations that would rule its mechanisms out. The solar-fold specification carries date priority in the programme record: a false-colour wavelength scan of the limb must show the disk grow, partially vanish in a caustic fold at the 4400–5000 K temperature-minimum band, and reappear — the standard many-layer reading predicts no fold, and the protocol runs on existing narrowband imagery. Internal gates bind with the same force: the W+1 rule dies if Weff departs from the integer 3 beyond 3σ, and the closure-schedule completion law (Section Paper IV) is pre-registered as a single-pass computation of the whole docking-lock curve — if it needs re-tuning at any rung, that is the finding.
5. Operating rules of the programme
(i) Pre-registration before instrumentation. (ii) Instruments validated on known answers before use. (iii) Whole-range, single-pass comparisons — no per-item re-reading. (iv) Convergent results claimed as convergent, shared-form results as shared form, never as discrimination. (v) Verdicts retained: the rotation-curve failure, the cubic-distance exclusion, and the withdrawn frustration selector all remain on display. (vi) Counts-indexed narration is inadmissible as evidence (the NP24-A rule): a regularity that merely re-reads the ledger cannot pass as a prediction.
6. Certification
| Item | Status |
|---|---|
| Specification corpus E13–E109 | 97 numbered specs; format of eq. (1) |
| Convergent anchors | labelled — no discriminating weight claimed |
| Divergent set | the framework's live exposure (§3) |
| Kill conditions | registered in advance (§4), incl. internal gates |
| Retired discriminants | shared-form entries removed and recorded (e.g. fragment-energy fine-tune) |
Methodological declaration. Rival frameworks appear in this programme exactly where they belong: as the named alternative in each discriminant row. No rival machinery is used to generate any SDT prediction. The programme's authority rests on its retained verdicts — the failures displayed beside the passes — and on the pre-registration discipline that prevents a prediction from being written after its answer.
References
- Experiments/E13.md … E109.md; Experiments/README.md.
- GATEWAY_PROCEDURAL.md — pre-registration and instrument-validation rules.
- Investigations/ — the executed subset with verdicts (NP27, GD05, CR12, OP07 protocol, TD03).
- Practice papers of this series: 15 · Optics & Photonics (the falsifier capstone) and each domain paper's certification table.