D-0011 · What counts as the same word
When are two written words the same word?
Two levels, and every match says which it reached. Nikud, te'amim, punctuation and final letters are always set aside. First, letter for letter (an exact match); only if that fails, the words are compared with ו and י disregarded, for full and missing spelling (a spelling match).
| Status | Ruled by the owner |
| Set | 20 Tishrei 5787 · 2026-10-01 13:22 |
| Recorded in | held ee3b45e1 |
| Rests on | D-0000 |
| Affects | every word-match result, D-0010 |
| Would change if | A case where the spelling level calls two different words the same, or misses a real spelling variant. |
The options on the table
- Two levels, both recorded — letter for letter first, then ignoring ו/י spelling; say which matched (chosen)
- Spelling differences ignored only (the machine's earlier guess)
- Letter for letter only
- Spelling differences ignored, plus a list of known substitutions (ה׳ for השם, abbreviations)
Why
The AI writes כוהן and the sefer has כהן; the AI writes ירושלים and the sefer has ירושלם. Letter-for-letter alone would miss real quotes; spelling-loose alone would never say how close a match was. Recording the level keeps both facts: "found, exactly" is stronger than "found, with spelling differences", and a reader can see which one it is. These are the "exact quote" and "normalised quote" of the record-first vocabulary (D-0004).
History
- 2026-10-01 04:45 — the machine's guess: spelling differences ignored only (experiments/2026-10-01-head-to-head).
- 2026-10-01 13:22 — ruled: two levels, both recorded.
What rests on this
Generated from the register (docs/decisions/0011-same-letters.md). A changed value is recorded as a new entry; the old one is never overwritten.