# n247 — The LIGO prediction: the glass dephasing (entry 39)

**Date:** 2026-07-12 · **Question (the human's): a prediction for LIGO?** Answer: the registry's only entry where LIGO can make a *positive discovery* rather than tighten a null — the vacuum's disorder read directly off waveform residuals.

## The prediction

GW phases random-walk on the vacuum's disorder: **Δφ = √(A·L)·2πf/c with A ≡ (δc/c)²·ℓ_d — one glass parameter.** The three-part in-data shape, all of which must hold together:

1. **√L across the catalog**: residual dephasing rms (after best-fit templates) grows as the square root of distance;
2. **∝ f within each event**: the merger dephases before the early inspiral — a frequency-linear phase noise no astrophysical systematic mimics;
3. **polarization-split**: the + and × contents dephase independently (the birefringent component) — testable on face-on/edge-on comparisons.

## The numbers

Self-consistency already sharpens the bound: clean template matches at Gpc through ~300 Hz require **A ≲ 2×10⁻¹⁶ m** (ℓ_d < 0.5 pm at the model's 2% per-cell contrast — one order below n246's estimate, which used the 100 Hz band). At that ceiling, the effect sits *just under* current single-event visibility — the best possible location for a prediction: **residual-stacking of N loud events probes A down by ~1/N** (10 events → 2×10⁻¹⁷ m; 100 → 2×10⁻¹⁸ m). O4/O5 catalogs decide territory the theory actually occupies.

## The recipe (for the analysis that could run today)

Stack post-template phase residuals of the loudest BBH events, binned by luminosity distance; fit Δφ²_rms = A·L·(2πf/c)²; report A or its upper limit. Cross-checks: the f-linearity within events; the polarization split; no correlation with detector state (the effect lives on the source-to-Earth path, not in the instrument).

## Stakes, both directions

- **Discovery**: a measured A with the three-part shape = the vacuum's disorder parameter, read at Gpc baselines — the glass observed. It would simultaneously price ℓ_d in the quench triangle (entry 38) and fix the GW-birefringence amplitude.
- **Nulls**: every clean catalog shrinks A — at ~2×10⁻¹⁸ m the KZ freeze-out correlation drops to 10⁻² pm at model contrast, squeezing the quench triangle from the GW side.
- **Kill**: a measured dephasing trend with the *wrong* shape — scaling ≠ √L, or frequency dependence ≠ linear, or no polarization component — kills the glass-disorder mechanism (not just bounds it).

**Class: shape [A-model — the random-walk form is the measured n222/n229 physics]; the contrast 2% [model number, declared — A is the honest parameter]; entry 39 registered [distinctive · pre-test]. LIGO's next thousand detections either find the vacuum's grain or grind it below a hundredth of a picometre.**