Phase 22.B Results: The Linear A / Linear B Morphological Gap Is Not a Length Artefact, but Within-Greek Variation Leaves Its Interpretation Unresolved.

Pre-registered results for Phase 22.B. The Linear A vs Linear B word-final -o gap survives stratification by word length in every tested class (Δ = 40.3, 47.2, 45.8 percentage points; p = 1e-4 throughout) and length-matched resampling (Δ_p05 = 37.97 against a threshold of 20). The gap is therefore not a word-length artefact. However, the pre-declared within-Greek yardstick control failed: Knossos-only and Pylos-only Mycenaean word types differ from each other by up to 30 percentage points on the same measure, so the Linear A / Linear B gap does not reach the pre-registered 2x margin over normal between-corpus variation in Greek. Verdict: SCALE-UNRESOLVED. Per the pre-declared consequence, the gap may not be cited as evidence of non-Greekness until a newly pre-registered Phase 22.C tests a range-non-overlap criterion. A post-hoc objection to the control's formulation is recorded but explicitly not acted upon. Pre-registration: doi 10.5281/zenodo.22742992.