# n70 — pre-registration: first heartbeat of tensor gravity (before code)

**The construction.** After n69, gravity's home is the tensor sector of
the link field: the components NOT constrained by the audit. Matter
couples to it through its SHAPE — stress, not consumption: an
anisotropic dressing prefers anisotropic link stiffness (a traceless
quadrupole source s_e on edges). Cost C(dk) = (1/2) dk^T G dk - s^T dk;
interaction between two sources: E_cross(d) = - s_A^T G^{-1} s_B.

**Why the sign should be gravity's, this time:** linear coupling to an
even/bosonic quadratic sector attracts LIKE sources (the spin-0/spin-2
exchange rule) — the same structural rule that made the audited
(constraint/odd) sector repel likes (n51 signed / n69 EM). If the
tensor channel attracts like stresses universally, gravity has a pulse
in this vacuum.

## Gates (2D, info-metric G, traceless quadrupole sources
## s = (+1 on x-edges, -1 on y-edges) at a site)
- H1 (SIGN): E_cross(d) < 0 for same-oriented quadrupoles at all d
  (joint cheaper than separate = attraction between like stresses).
- H2 (TENSOR STRUCTURE): orientation dependence — same-oriented (xx)
  vs crossed (xy) sources differ in sign or magnitude with a definite
  pattern (a scalar channel would be orientation-blind); report the
  ratio.
- H3 (RANGE, the honest one): measure the decay of E_cross(d).
  EXPECTATION DECLARED: SHORT-RANGED — the bare metric G is gapped
  (spectrum [0.023, 0.031], no pole), so this static channel cannot be
  1/r^2 gravity by itself. The masslessness must come from the
  substrate's acoustic TT phonons (omega -> 0) under monitoring — the
  year-1 mechanism, which is the registered NEXT construction (n71),
  not this one. H3 records the gap that n71 must close.
- KILL: H1 fails (like stresses repel in the tensor channel too) —
  then this vacuum has no attractive tensor sector and gravity has no
  home here at all: the reinterpretation's promissory note bounces.

## Honesty clauses
- The stress coupling s is INSERTED at this stage (the minimal linear
  source); deriving it from the dressing's shape is part of n71.
- 2D, L = 24; static; no claim of masslessness here (H3 exists to
  prevent exactly that overclaim).

## RESULTS (same day) — the kill clause fires
- H1 FAIL: same-oriented quadrupoles REPEL at every separation
  (E_xx > 0, d = 1..10); crossed ones attract with EXACT antisymmetry
  (E_xy = -E_xx to machine precision).
- H2 PASS, emphatically: perfect orientation-alternating tensor
  structure — but alternating is the OPPOSITE of universal. Gravity's
  defining property (everything attracts everything) is absent.
- H3: r^-3.5, short-ranged as declared (same exponent as the n69
  neutral residual — the static tensor channel IS the residual-force
  sector, not a fundamental one).
- STRUCTURAL DIAGNOSIS: the isotropic (monopole) source direction is
  the audited combination — spoken for by the electrostatic sector.
  What remains for a static tensor coupling is traceless quadrupole
  physics, which alternates by construction. NO STATIC CHANNEL OF THIS
  VACUUM CAN BE GRAVITY: the audited scalar is EM (n69), the free
  tensor is an orientation-alternating residual (this file), the
  energy route is concave (n45), the information route was EM's sign
  all along.
- THE ONE DOOR LEFT, and the circle closes: gravity's only remaining
  home in this framework is the DYNAMICAL mechanism — attraction
  generated by continuous MONITORING of the substrate's massless TT
  phonons: Bath-TT, year 1's original idea, the theory this program
  was founded to test and, in two years of killing everything else,
  never actually ran in a toy. n71 is that experiment. Every other
  door has been closed by a measurement with its name on it.