ORAL AUDIT — gen_17_boarding
python3 logic/solo_tools/verify_claims.py logic/oral_audit/manifests/gen_17_boarding_claims.jsonClaims manifest (verified live at render time)
| status | id | source | claim · evidence |
|---|---|---|---|
| VERIFIED | G17-01 | Minchat Shai on Gen 7:2 (Ramah; corrected Masorah) | הטהורה ("the pure one," with the article) — the Ramah: plene with vav exactly TWICE in the Torah, "both verse-openings" — the boarding commands 7:2 and 7:8; every other pure-form defective (Minchat Shai calls our printed Masorah here "words of error"). SNAPSHOT holds THREE plene tokens: the Ramah's verse-initial pair AND 8:20 (the altar verse) — the third is a deviation from his rule, banked as a discrepancy for the gen_20 audit. (The pure LAMPSTAND's tokens and Lev 20:25, same pronunciation, are all defective — consistent with the rule.) full=['Gen 7:2', 'Gen 7:8', 'Gen 8:20'] of 7 tokens |
| VERIFIED | G17-02 | Minchat Shai on Gen 7:8 (Pesachim 3a; Bereshit Rabbah 32) | DISCREPANCY OBSERVATION, the tradition's own — the Talmud: "the verse BENT EIGHT LETTERS rather than utter an ugly word" (אשר איננה טהרה, "which is not pure," for הטמאה, "the impure"). Minchat Shai's challenge: טהרה here is written DEFECTIVE (four letters), so the detour is SEVEN letters, not eight — and he resolves it by Bereshit Rabbah's version, which counts WORDS ("two and three words"), not letters. SNAPSHOT sides with his arithmetic: the defective spelling locked. letters=טהרה vs טהרה |
| VERIFIED | G17-03 | Minchat Shai on Gen 7:11 (Hilleli vs Yerushalmi; Ramah; Or Torah) | DISCREPANCY OBSERVATION — מעינות ("fountains"): "a great dispute on this word." The Hilleli codex, the Ramah, and Or Torah rule 7:11 PLENE — "full when the fountains were full" — and 8:2's closing token defective, "drained when they drained": the spelling-homily rides the ruling. SNAPSHOT writes BOTH tokens defective (מעינת) — the Yerushalmi camp — against the Ramah's ruling and its homily. Both camps recorded; the check locks zero plene tokens. full=[] of 2 tokens |
| VERIFIED | G17-04 | Kitzur Baal HaTurim on Gen 7:4 (Masorah) | היקום ("the standing-substance") — the Kitzur's Masorah trio, all THREE in the Torah: "and I will wipe out all the standing" (7:4), "and He wiped out all the standing" (7:23), "and all the standing at their feet" (Deut 11:6 — Korach's household): "as the flood generation sinned from too much good, so Korach from his wealth." SNAPSHOT: exactly three tokens. count=3 refs=['Deut 11:6', 'Gen 7:23', 'Gen 7:4'] |
| VERIFIED | G17-05 | Kitzur Baal HaTurim on Gen 7:16 (Masorah) | והבאים ("and those coming") — the Kitzur's Masorah pair spans out of the Torah: the boarding animals (7:16) and "those who came with Zerubbabel" (the return list): "as here they came OF THEMSELVES, so there they went up of themselves." Torah-side: this is the word's only token — and the pair is an inheritance edge (Torah → Writings) of exactly the class the edge layer records. count=1 refs=['Gen 7:16'] |
| VERIFIED | G17-06 | Minchat Shai on Gen 7:6 | ונח בן שש מאות שנה ("and Noah was six hundred years old") — שש with its own accent and NO maqqef (joining hyphen): Noah's age stands unjoined. SNAPSHOT: no maqqef after שש. maqqef_after=0 |
| VERIFIED | G17-07 | Minchat Shai on Gen 7:11 | בשנת שש מאות שנה ("in the six-hundredth year," the date formula) — here שש WITH maqqef and no lengthening mark: the same number, joined when it dates the flood and free when it ages the man — Minchat Shai notes both pointings five verses apart. SNAPSHOT: maqqef after שש. maqqef_after=1 |
| VERIFIED | G17-08 | Minchat Shai on Gen 7:11 | בחדש השני ("in the second month") — zaqef qatan ("small upright" pause accent) on the nun. SNAPSHOT agrees. mark=zaqef_qatan |
| VERIFIED | G17-09 | Minchat Shai on Gen 7:12 | ויהי הגשם על הארץ ("and the rain was upon the earth") — על joined by maqqef (joining hyphen). SNAPSHOT agrees. maqqef_after=1 |
FOLDED INTO THE UNIT same day (one source of truth). Block 1 of the second five-block run — FIRST audit carrying the INHERITANCE EDGES section (owner design session 2026-08-08: edges = typed, attested links from export_links; Torah side verified, inheritor wording pending the Prophets/Writings instrument). Zero web fetches. Manifest: logic/oral_audit/manifests/gen_17_boarding_claims.json — 9 VERIFIED · 0 FAILED · 0 UNCHECKABLE, including TWO discrepancy observations — the richest discrepancy block yet.
Sources: Minchat Shai on Gen 7 (12 notes; the 7:23 notes banked for gen_18) · Kitzur Baal HaTurim on Gen 7 (7 notes) · Rashi on Gen 7 (skimmed; 7:23 grammar note banked) · span map (Bereshit Rabbah 32).
CROWNS (chain-attested + DB-verified)
- G17-01 the pure-word's two openings (+1): הטהורה ("the pure one") plene exactly at the Ramah's verse-initial pair — the boarding commands 7:2 and 7:8 — while the pure LAMPSTAND's tokens stay defective; Minchat Shai calls our printed Masorah here "words of error" and SNAPSHOT sides with his correction, EXCEPT one token: 8:20 (the altar verse) is plene against the Ramah's rule — banked as a discrepancy for the gen_20 audit.
- G17-04 the standing-substance trio: היקום — exactly three Torah tokens: the wipe-threat (7:4), the wipe-report (7:23), and Korach's household (Deut 11:6): "as the flood generation sinned from too much good, so Korach from his wealth" (Kitzur).
- G17-05 those who came of themselves: והבאים — the word's only Torah token (the boarding animals), paired by the Masorah with "those who came with Zerubbabel": they came of their own accord, both times (Kitzur).
- G17-06/07 the two sixes: שש ("six") — free-standing with its own accent when it AGES Noah (7:6), maqqef-joined (joining hyphen) when it DATES the flood (7:11): Minchat Shai notes both pointings five verses apart; SNAPSHOT's maqqef column confirms the contrast.
- G17-08/09 zaqef qatan on the second-month word (7:11); maqqef on על־הארץ ("upon the earth," 7:12) — both as ruled.
DISCREPANCIES
- G17-02 the seven bent letters (the tradition's own): Pesachim 3a — "the verse bent EIGHT letters rather than utter an ugly word" (אשר איננה טהרה, "which is not pure," in place of הטמאה, "the impure"). Minchat Shai's challenge: the pure-word here is DEFECTIVE — four letters — so the detour is SEVEN, and he resolves by Bereshit Rabbah's version, which counts WORDS. SNAPSHOT sides with his arithmetic: the defective spelling locked, the Talmud's count works only on the plene spelling SNAPSHOT does not have.
- G17-03 the fountains that did not fill: מעינות ("fountains," 7:11) — "a great dispute": Hilleli, the Ramah, and Or Torah rule the breach-verse PLENE ("full when full") and the closing verse (8:2) defective ("drained when drained") — a spelling-homily. SNAPSHOT writes BOTH tokens defective — the Yerushalmi camp — against the ruling, and the homily dies at the letter. Both camps recorded.
INHERITANCE EDGES (attested in export_links; Torah side verified)
- Gen 7:2 → Job 12:10 — the clean/unclean boarding command linked to "in whose hand is the life of every living thing": edge attested; inheritor wording pending the Prophets/Writings instrument.
- Gen 7:11 → II Samuel 1:21 — the heavens' windows breached, linked to David's Gilboa curse (no dew, no rain): edge attested; inheritor wording pending.
- (G17-05's Zerubbabel pair is a third edge, carried as a verified claim because the Kitzur names it.)
READINGS FILED (named-only, unclaimed)
"Thee have I seen righteous" — but not "righteous and wholehearted": part of a man's praise in his presence (Eruvin 18b class, via Rashi 7:1 — attests the unit's held praise-abridgement delta on this-generation) · the seven days = mourning for Methuselah (Kitzur 7:10; letter-arithmetic class) · the seventeenth = "they spend their days in GOOD" — the good-word's value (Kitzur 7:11; same class) · the two stars taken from the Pleiades to breach the windows (Kitzur 7:11, the Berakhot 59a tradition) · the Lord shut him in — bears and lions guarded the door (Bereshit Rabbah 32:8 via Rashi 7:16 — beside the unit's sealed-door close) · the first rain fell gently, a chance for repentance (Rashi 7:12).
TIMING + PROCESS NOTES
BLOCK CYCLE — 4 minutes 17 seconds (13:38:26-13:42:43; the run was interrupted once before the span queries — pre-interrupt work not counted). One verifier iteration: G17-01's translit filter caught seven same-sounding tokens (the pure lampstand among them) — the spelling filter, not the sound filter, is the right census tool; the fix ENRICHED the claim (the lampstand's consistent defectiveness). EDGES section debuted at near-zero cost: one query, two attested edges recorded with pending status.