Grounding Exercise: M5 Gravity — GW170817 Gate + Poland Fiber Program

Cell under grounding: FORCES / Gravity / Liquid Crystal (M5) in MODELS.md. Current status: ⚠️ — coupling mechanism measured (GEM ∝ (b·g)², zero at zero boost, m5_8_2q_delta_scaling.py), dynamical metric not implemented; the de Broglie clock sector IS implemented and validated. This document: the GW170817 radiative gate (§0–6), the terrestrial extension — desk gates on existing data plus the Poland fiber experiments (§7–10) — and the priority logic that orders them by how much model-unknown each uniquely grounds.

0. The radiative claim, split into a gate

GW170817/GRB 170817A gates any gravity sector with two clauses:

G-a (existence): the model's gravity sector supports a radiative
    (wave) mode — GWs demonstrably exist;
G-b (speed):     that mode propagates at c within
                 −3×10⁻¹⁵ ≤ c_g/c − 1 ≤ +7×10⁻¹⁶.

A model can fail G-a without reaching G-b: the split converts "compare with GW170817" from a number-check into a structural audit.

1. Pre-registration atom (created FIRST, hash-pinned)

atom PREREG { uri: file://.../gates/gw170817_prereg.md,
              digest, t, via: tool://git@<commit> }

Frozen before any run: (1) MODE IDENTIFICATION — which excitation of the boost/tilt sector counts as gravitational radiation, distinguished from the EM tilt modes of m5_6_4a (this choice is the tuning surface, so it locks first); (2) measurement protocol — dispersion fit ω(k) at 24³/32³/48³, speed at small k, convergence tolerance; (3) pass criteria — G-a: finite-energy radiative mode; G-b: |c_g/c − 1| ≤ 1×10⁻³ in-platform (declared numerical floor, physical bound recorded as target); (4) calibration freeze by digest. The lock protects MODEL-SIDE choices: the famous observational answer cannot be quarantined from authors, but the future self who would pick the mode that comes out at c can be locked out.

2. Atom ledger (radiative gate)

GW-STRAIN    https://gwosc.org/eventapi/.../GW170817 (HDF5, digest)
GRB-TIME     doi://10.3847/2041-8213/aa920c   Δt = 1.74 ± 0.05 s
KILONOVA-IMG doi://10.1126/science.aap9811    (Swope image — host
             anchor provenance, never a fitted input)
DISTANCE     doi://10.3847/2041-8213/aa920c   D ≈ 40 Mpc
M5-CODE      tool://openwave@<commit>
M5-GEM       file://.../m5_8_2q_delta_scaling.py
M5-EMWAVE    file://.../m5_6_4a_hydro_em.py   (anti-conflation)
CALIB        file://.../gates/unit_chain_v1.md
ENV          tool://python-env@<hermetic-digest>

3. Plan (SSA form)

s0 := load(PREREG)                        s5 := disp(s4, 24³)
s1 := load(M5-CODE)                       s6 := disp(s4, 32³)
s2 := load(CALIB)                         s7 := disp(s4, 48³)
s3 := load(ENV)                           s8 := speed(s5,s6,s7,s2)
s4 := identify_mode(s1, s0.mode_spec)     s9 := load(GRB-TIME)
s10 := compare(s8, s9, s0.criteria)       s11 := emit_certificate

Taint: obs://gw170817 atoms are sink-only until s9 — barred from identify_mode and speed extraction.

4. Automaton

START → PREREGISTERED → MODE_BUILT → SPEED_MEASURED → GATED
⊥ on: compare before prereg; G-b without G-a provenance;
identify_mode reading obs:// atoms; calibration newer than prereg;
status write without certificate.

5. Certificate → cell

{ plan-hash, PREREG digest, DAG s0..s11 with tool digests,
  dispersion tables + convergence, verdicts G-a/G-b,
  append-only trace position }
Cell: ⚠️/✅/❌/🚧 [certified] — verdict — cert:<hash>
Replay: fresh machine ⇒ identical digests. Injection: perturb
CALIB one digit ⇒ exactly s8,s10,s11 dirty.

6. Honest expected outcome (radiative gate)

Likeliest first certificate: G-a blocked(missing-tool) — no constructor for a boost-sector radiative excitation distinct from EM tilt; certificate names the unblocker (TOOL-D: wave launcher + energy-flux meter). Cell: 🚧 [certified blocked]. Both live outcomes pre-wired: certified first gravity ✅, or certified ❌ with dispersion tables — a publishable negative.

TERRESTRIAL EXTENSION: EXISTING DATA + THE POLAND FIBER PROGRAM

The GW gate needs the sector M5 lacks (radiation). The terrestrial gates aim at the sector M5 HAS: the validated de Broglie clock and the measured (b·g)² coupling. Strategy: ground against archival data first (desk gates, zero cost), then buy new data only where the unknown is one no archive reaches. Each gate names the smallest model-side tool it requires; priority = archival leverage first, then unknowns per złoty.

7. The model-side tools (smallest first — this ordering IS the plan)

TOOL-A  static clock-shift law: boost-sector response to a static
        source → clock rate r(φ, g). Does M5's clock couple to
        POTENTIAL φ (GR-like) or to FIELD MAGNITUDE g (the naive
        reading of GEM ∝ (b·g)²)? No dynamical metric needed —
        static weak-field response only. Smallest tool, biggest
        gate coverage (D1, D2 partially, F1).
TOOL-B  GEM loop-integral / frame-dragging analog: the model's
        Lense–Thirring prediction (D3, F2).
TOOL-C  medium admittance: response amplitude+phase to a moving
        mass at 10³–10⁷ m scales (D2, F3).
TOOL-D  radiative launcher (§6) — the big one; last.

8. Desk gates on EXISTING data (run before any trench is dug)

D1 — Geoid clock gate. Sea level is an equipotential: between Singapore-latitude and Helsinki-latitude labs, g differs 0.5% while sea-level clock rates agree to first order. Atoms: published inter-NMI fiber/satellite clock comparisons + Tokyo Skytree 450 m campaign (doi atoms). Requires TOOL-A. Outcomes: · M5 derives φ-coupling → passes decades of archival timekeeping → clock-redshift sub-criterion certified ✅ against existing data; · M5 derives g-coupling or mixed with |g-part| ≳ 10⁻⁵ of total → certified ❌ from archives alone — no experiment needed; · TOOL-A not derivable → certified 🚧 naming TOOL-A (and note: this tool is far smaller than the metric; a certified 🚧 here is a concrete, sized work item, not a shrug).

D2 — IGETS tidal admittance. Free superconducting-gravimeter archives, 10⁻¹¹ g, decades: amplitude ratio + phase lag of the solid- Earth tide vs Newtonian prediction. Requires TOOL-C (static limit via TOOL-A first). Outcomes: admittance within data error → medium stiffness/lag certified consistent at 10⁴–10⁸ m; predicted phase lag outside → ❌ with the transfer function attached; no prediction → 🚧 naming TOOL-C.

D3 — Frame-dragging bound. GP-B (geodetic 0.3%) + LARES. M5's gravity IS gravitomagnetic-flavored, so this is the archival gate aimed at its actual sector. Requires TOOL-B. A GEM sector predicting Lense–Thirring off by more than the LARES error bar dies at a desk.

9. The Poland fiber gates (new data, sited on OPTIME/PIONIER)

Assets already in place: KL FAMO Toruń (optical lattice clocks), GUM Warsaw + AOS Borowiec (TAI time labs), OPTIME fiber time-transfer backbone, Książ geodynamic laboratory (quiet tunnels, tiltmeters).

F1 — Mine clock line: KGHM Polkowice–Sieroszowice, surface ↔ ~1000 m level (alt: Kasprowy Wierch ↔ Zakopane, +1100 m). THE UNKNOWN IT UNIQUELY GROUNDS: every archival comparison (D1) lives above the surface, where φ and g co-vary one way along altitude; underground the gradient regime changes sign — φ and g DECOUPLE. If D1 leaves any mixed-coupling window open, F1 closes it: a clock law r(φ, g) that mimics GR above ground cannot also mimic it below. Atoms: KGHM shaft survey (Δh, density logs → φ(depth) model), OPTIME link calibration, clock certificates from Toruń; taint rule: the mine's φ/g profile atoms are sink-only relative to the model run. Outcome tree: match GR and M5-with-TOOL-A → strongest terrestrial ✅ available to the clock sector; deviation at declared sensitivity → the headline; TOOL-A absent → the gate WAITS (F1 is pointless before D1 — priority discipline).

F3 — Porąbka-Żar scheduled-mass admittance. Upper reservoir: ~2×10⁹ kg cycled ~430 m daily on operator logs; powerhouse in-mountain. THE UNKNOWN: D2's tides drive at two fixed frequencies and planetary geometry; Żar is a CONTROLLED source — different geometry, schedule you can pre-register against, and the pumping log is a source atom exactly like the dam/dewar pattern. Instruments: one gravimeter + strain fiber up the penstock right-of-way. Grounds TOOL-C's transfer function where the drive is known to the kilogram-second.

F2 — Żar (or Książ) fiber loop: the one-form gate. Closed-path clock transport integral must vanish in a static field iff the clock field is curl-free; the only physical residue is gravitomagnetic. Fiber gyros cannot reach Lense–Thirring (ring-laser territory), so F2's value is structural: M5 must PRE-REGISTER its loop integral (TOOL-B); a large predicted curl already died at D3, a GR-size one makes the loop a bounded-null replay experiment under Żar's moving mass. Cheapest to add (dark-fiber ring), last in priority.

10. Priority order, and why

P0  D1 → D2 → D3   zero cost; three certified verdicts (✅/❌/🚧
                    each) grounding M5 against decades of data;
                    forces TOOL-A/B/C to be built or named.
P1  F1 (mine line)  first new data: the φ/g decoupling regime no
                    archive contains — the sharpest question for a
                    boost-sector clock, aimed at M5's validated
                    sector, on infrastructure Poland already owns.
P2  F3 (Żar mass)   controlled-source admittance: grounds dynamics
                    without needing radiation; shares TOOL-C with
                    D2 so the model work amortizes.
P3  F2 (loop)       structural pre-registration + bounded null;
                    rides F3's site.
P4  GW gate (§0–6)  waits on TOOL-D, the largest tool — but every
                    earlier certificate narrows what TOOL-D must
                    produce, and D1's φ-coupling verdict is a
                    boundary condition any future metric must obey.

The grounding logic in one line: EXISTING data adjudicates every claim the model can currently compute (and certifies exactly which tools are missing when it cannot); the fiber experiments are purchased in the order of the unknowns archives cannot reach — underground φ/g decoupling first, controlled-source dynamics second, curl structure third, radiation last. Every certificate, pass or fail or blocked, lands in the append-only ledger; per evidence_loss, none of them can ever be quietly unsaid.

chadbrewbaker/Polish_gravity.md

Select an option

No results found