source_gravity_admissibility

Admissible source processes with identical clocks and different transport

The proof strengthens the previous two-field comparison: these are actual words in the existing coherent twelve-port source, whose A1--A3 membership was constructed previously. One stationary source and the entire resting-clock channel are shared. Two fixed local coin policies give different finite receiver probabilities. The missing spatial response is a function of source strength; clock calibration and the typed A3 state objective do not select it.

This is a non-entailment result, not a derivation of physical gravity, Maxwell photons or selection by the integer-repair simulator. RG and CCG remain proposed services in the named mathematical source. The contract distinguishes the exact finite exit from the long-wavelength interpretation and physical parent obligations.

From the repository root:

python -m source_gravity_admissibility.verify
python -m pytest -q code/source_gravity_admissibility

Set PYTHONPATH=code first. The checked-in receipt.json contains the complete compact experiment census. To reproduce it:

python -m source_gravity_admissibility.build

The producer solves the grounded source equations exactly and executes six-level native pulses reconstructed from the source response. The checker imports neither that solver nor its circuit constructor: it checks local source equations, independently reconstructs the response, and uses Clifford projectors to build every reflected flight. It also replays the retained native basis words on every input column. The finite click gap and error intervals use exact rational arithmetic; floating-point matrices are corroborating executions, not interval proofs or experimental data. No external measurements are newly consumed.

The primary experiment uses five sites and all three Clifford flight bases, four source strengths and both policies. Eight additional 27-site three-dimensional executions retain massive and massless sectors. Complete clock matrices and small-spacing Fourier matrices test the channel and principal-symbol calculations. The exact one-axis massive dispersion is also checked at 36 parameter settings with all four bands; the velocity freedom is present before a continuum approximation. An inhomogeneous native-word regression also checks the required Hermitian gradient term and the quadratic local expansion error in all three Clifford directions; removing that term fails the control. Only the named receiver read is a local physical report; full-matrix comparisons are software checks, not an observer pooling spatial quantum states instantaneously.