Pre-registration: Is the Linear A / Linear B Morphological Gap a Word-Length Artefact? (Phase 22.B).

Locked pre-registration for Phase 22.B. Phase 22 showed that length-matched resampling can reverse the sign of a word-final vowel comparison between corpora with different word-length profiles. This phase re-tests the project's strongest prior finding — the Linear A vs Linear B word-final -o gap (8.7% vs 49.6%) — stratified by word length, with a within-Greek yardstick control. Hypotheses, metrics, thresholds and the full verdict table, including the pre-declared retraction condition, were fixed before any test statistic was computed. Deposit includes the deterministic runner, SHA-256 checksums and OpenTimestamps proofs. No results. LOCK STATEMENT: this record's publication date and the accompanying OpenTimestamps proofs constitute the lock; the status line inside the .md file predates deposit and was left unedited so the proofs remain valid. SHA-256 of 170_phase22B_prereg_v1.md: 506b997adb2ae54a16bc52be13e489a7a50185b4be0752f66ce8c0b39f23819b. Version v2 (2026-09-15): OpenTimestamps proofs upgraded to confirmed Bitcoin block attestations. The pre-registration and runner are anchored in blocks 966927, 966928 and 966971, so timestamp priority is independently verifiable without reliance on any calendar server. Adds 170_phase22B_prereg_v1.PROOF.txt recording SHA-256 checksums, DOIs and block heights. No change to the pre-registration text, the runner, or the verdict.