HermitCrab analysis: query lexical reachability before materializing compounding splits (rejected: ceiling about 2% of clones, 9% of traversals) #585

The idea was to check lexical reachability before AnalysisCompoundingRule materializes its splits. It was not built.

Measured. A census on the heaviest Sena probe counted 35,176 splits, 5,524 root candidates and 1,121 outputs. That work is about 2% of word clones and 9% of FST traversals.

Why rejected. The ceiling is too small for the added indexing and correctness surface. Real compounding also prevents the simpler grammar-wide lexical gate. Reopen only if another grammar shows split construction as a dominant measured bucket and supplies a sound, stratum-aware reachability predicate.

Record: https://github.com/sillsdev/machine/blob/2eba2f502338705a4d529d7034cecb6303db7f35/docs/rejected-optimizations/compounding-split-prefilter.md
Ledger: https://github.com/sillsdev/machine/blob/2eba2f502338705a4d529d7034cecb6303db7f35/docs/hermitcrab-optimization-ledger.md; disposition: https://github.com/sillsdev/machine/blob/2eba2f502338705a4d529d7034cecb6303db7f35/docs/hermitcrab-perf-2026-09-disposition.md
From #490 (closed).