# n139 — the preferred-frame gauntlet: alpha_1, alpha_2
# (2026-07-11. The confrontation nobody had run: every preferred-frame
#  gravity in history dies here. Ours has a rest frame by construction.)

## The exposure
The substrate has a rest frame (n118: ordering fundamental, Lorentz
IR-emergent); the Solar System moves through it at w ~ 370 km/s. Any
O(1) coefficient on (w/c)^2 in the two-body statics produces
alpha_2-class effects ~ 1.5e-6 against the solar-spin-alignment
bound 1.6e-9: death by three orders. This is the classic execution
ground of preferred-frame gravities (Nordtvedt's blade).

## The computation, and the escape that was already built
The audit-mediated interaction of a comoving pair is filtered by the
monitor's response at the Doppler frequency omega = k.w. The
overdamped filter's corrections go as (k.w/gamma)^2 — NOT (w/c)^2:
   solar system: (k.w/gamma)^2 = 1.9e-49  (margin: 8e39)
   binary pulsars:               2.8e-45  (same verdict)
THE HIERARCHY IS THE PROTECTION. The audit responds at gamma =
5.7e18 s^-1; no astrophysical motion can Doppler it. The same
overdamping Q that makes gravity weak (n128: the two-channel ratio)
makes it frame-blind. One structure now carries four jobs: the
weakness of gravity, the CPS filter (n125), the CMB-frame
insensitivity of the statics, and the knee's location.

## PREDICTION 29 — the preferred frame stays hidden in gravity
alpha_1 and alpha_2 are zero at every reachable precision, forever —
not because there is no preferred frame (there is one), but because
the monitor is 40 orders too fast to reveal it in the force sector.
LLR, pulsar timing, solar-spin analyses keep returning null.
KILL: any confirmed preferred-frame PPN detection — it would
simultaneously falsify GR and this framework, in opposite
directions (GR has no frame to show; we have one that must stay
hidden by the measured margin).

## The cost the same computation names (carried openly)
If frame effects enter at (k.w/gamma)^2, so do retardation effects:
the mediator filter CANNOT supply the Einstein-Infeld-Hoffmann
velocity terms at O(v^2/c^2) — and those are real, measured physics
(the PN structure of binary-pulsar timing). In this framework they
must come from the COVARIANCE of the emergent sector (the statics
rewritten on the cone), not from the mediator's dynamics. This is
not a new debt — it is the tensor junction's debt, sharpened to a
precise statement: THE JUNCTION MUST DELIVER EIH, and the mediator
cannot help it. The no-retreat list (quadrupole, factor 4) gains a
third item with existing experimental executioners.