ORAL AUDIT — gen_09_helper_woman_first_speech
python3 logic/solo_tools/verify_claims.py logic/oral_audit/manifests/gen_09_helper_woman_first_speech_claims.jsonClaims manifest (verified live at render time)
| status | id | source | claim · evidence |
|---|---|---|---|
| VERIFIED | G09-01 | Kitzur Baal HaTurim on Gen 2:21 (Masorah pair) | va-yisgor ('and He closed') occurs exactly twice in the TORAH: Gen 2:21 (the flesh over Adam's side) and Gen 7:16 (the ark door behind Noah) - the Torah's only two closings, both divine; the Kitzur's Masorah pairs exactly these two (later-Prophets human closings lie outside the Torah instrument) count=2 refs=['Gen 2:21', 'Gen 7:16'] |
| VERIFIED | G09-02 | Kitzur Baal HaTurim on Gen 2:20 (Masorah, pair Eccl 2:21) | u-le-adam ('and for the man') occurs three times in Scripture; the two Torah tokens are Gen 2:20 (no helper found) and Gen 3:17 (the sentencing: 'because you listened to your wife') count=2 refs=['Gen 2:20', 'Gen 3:17'] |
| VERIFIED | G09-03 | Kitzur Baal HaTurim on Gen 2:21 (Masorah, pair Ps 78:28) | va-yapel ('and He caused to fall') occurs three times in Scripture; the two Torah tokens are Gen 2:21 (the deep sleep) and Num 35:23 (the manslaughter fall) count=2 refs=['Gen 2:21', 'Num 35:23'] |
| VERIFIED | G09-04 | Kitzur Baal HaTurim on Gen 2:22 (Masorah); Minchat Shai on Gen 2:22 | va-yevie-ha ('and He brought her') has exactly two Torah tokens: Gen 2:22 (the woman to Adam) and Gen 24:67 (Rebekah into Sarah's tent by Isaac) - Adam's bride and Isaac's bride carried by the same word count=2 refs=['Gen 24:67', 'Gen 2:22'] |
| VERIFIED | G09-05 | Minchat Shai on Gen 2:22 ('two, missing yod; its fellow: Isaac into the tent') | Both va-yevie-ha tokens are written DEFECTIVE (missing the yod) - zero full spellings in the Torah full=[] of 2 tokens |
| VERIFIED | G09-06 | Kitzur Baal HaTurim on Gen 2:23 (Masorah, pair Job 37:1) | le-zot ('to this one shall be called woman') occurs twice in Scripture; in the Torah only Gen 2:23 count=1 refs=['Gen 2:23'] |
| VERIFIED | G09-07 | Minchat Shai on Gen 2:25 (Masorah, pair Job 22:6) | arumim ('naked', 2:25) is written FULL with the vav - the Masorah pairs it Scripture-wide with Job 22:6 and marks the full-and-dagesh spelling as the naked-class against the cunning-class (arum, 3:1) full=['Gen 2:25'] of 1 tokens |
| VERIFIED | G09-08 | Minchat Shai on Gen 2:25 ('defective, no vav in the middle, in all books') | yitboshashu ('they were not ashamed') is written DEFECTIVE - no second vav full=[] of 1 tokens |
| VERIFIED | G09-09 | Rashi on Gen 2:21 (citing Exod 26:20, the tabernacle side); Eruvin 18a (side/face debate) | tzela ('side/rib') occurs 21 times in the Torah: the two Eden tokens (2:21 the taking, 2:22 the building of the woman) and nineteen sanctuary tokens - the ark's and tabernacle's sides. Outside Eden, every Torah tzela is a sanctuary side count=21 refs=['Exod 25:12', 'Exod 25:14', 'Exod 26:20', 'Exod 26:26', 'Exod 26:27', 'Exod 26:35', 'Exod 27:7', 'Exod 30:4', 'Exod 36:25', 'Exod 36:31', 'Exod 36:32', 'Exod 37:27'] |
| UNCHECKABLE | G09-10 | Minchat Shai on Gen 2:21 ('a paseq stands between the two words, for the honor of the Name') | A paseq (separator bar) stands between Elohim ('God') and tardemah ('deep sleep'), keeping the Name apart from the sleep-word instrument gap - SNAPSHOT encodes zero paseq marks corpus-wide (verified); claim stands on Minchat Shai's authority only |
| UNCHECKABLE | G09-11 | Minchat Shai on Gen 2:22 (ha-tzela pointing: qamatz vs patach manuscript dispute) | Manuscript traditions dispute the pointing of ha-tzela ('the side') - qamatz in the precise codices vs patach in most prints instrument gap - SNAPSHOT carries one pointing tradition; manuscript variants not encoded |
FOLDED INTO THE UNIT same day (standing owner order: one source of truth). The unit yaml's oral_audit_note_en block is canonical for the crowns; this file is the audit RECORD only.
Second retro audit; FIRST FULLY LOCAL-FIRST RUN — zero web fetches (all sources read from Data/sefaria_export/ via export_texts + export_links in derivation.sqlite). Manifest: logic/oral_audit/ manifests/gen_09_helper_woman_first_speech_claims.json — verifier: 9 VERIFIED · 0 FAILED · 2 UNCHECKABLE.
Sources read locally: Kitzur Baal HaTurim on Gen 2:18-25 (6 notes, verbatim Hebrew + English) · Minchat Shai on Gen 2:21-25 (7 notes, verbatim) · Rashi on Gen 2:18-25 (16 notes with his Talmud loci named) · export_links span map (Bereshit Rabbah ×28, Yevamot, Eruvin, Berakhot, Sanhedrin, Niddah, Ketubot loci per verse).
CROWNS (chain-attested + DB-verified)
NEW:
- G09-01 the two divine closings: va-yisgor ("and He closed") occurs exactly twice in Scripture — the flesh over Adam's side (2:21) and the ark door behind Noah (7:16). Both Torah; identical consonants. Kitzur Baal HaTurim.
- G09-04/05 the two brides: va-yevie-ha ("and He/he brought her") — exactly two Torah tokens: the woman brought to Adam (2:22) and Rebekah brought into Sarah's tent (24:67); BOTH written defective (missing yod), per Minchat Shai's exact note.
- G09-02 the helper frame: u-le-adam ("and for the man") — two Torah tokens: "no helper found" (2:20) and the sentencing "because you listened to your wife" (3:17).
- G09-03 the two falls: va-yapel ("and He caused to fall") — two Torah tokens: the deep sleep (2:21) and the manslaughter fall (Num 35:23).
- G09-06 le-zot ("to this one") Torah-unique (pair: Job 37:1).
- G09-07/08 the naked spellings: arumim ("naked") written FULL (Masorah pair Job 22:6) while yitboshashu ("they were not ashamed") is written DEFECTIVE — Minchat Shai, both letter-exact in SNAPSHOT.
ATTESTING the frozen unit:
- G09-09 the unit's own note that the side-noun recurs ×19 in the sanctuary is Rashi's cross-reference (Exod 26:20) completed: tzela ("side") = 21 Torah tokens — 2 Eden + 19 ark/tabernacle sides; outside Eden every tzela is a sanctuary side. The woman and the sanctuary are the only things built from one.
- The unit's sagar-debut claim gains the Masorah pair (G09-01); its naked/cunning bridge gains the Masoretic spelling side (G09-07); its boosh-debut-negated gains the defective spelling (G09-08); its named-only Yevamot 63a pointer is now read and linked locally.
READINGS FILED (named-only, unclaimed — triage law)
Yevamot 63a (worthy = a help, unworthy = opposed; the failed search among the creatures) · Eruvin 18a (two-faces; side-or-tail dispute) · Berakhot 61a (built like a storehouse, narrow above wide below; He braided her hair) · Sanhedrin 39a (He put him to sleep so she not be despised) · Sanhedrin 57b-58a: the Noachide forbidden-relations derived phrase-by-phrase from 2:24 — attesting this unit's PATTERN ("therefore"-generalization) read as LAW GRAMMAR, the second command-verse legal derivation after gen_08's Sanhedrin 56b · Niddah 45b (va-yiven, "and He built," read as binah, understanding — given to the woman in extra measure) · Bereshit Rabbah 17-18 (×28 links on the span) · Rashi/BR 18:4 (ish/isha sound-match proves the Holy Tongue at creation) · Kitzur on 2:24-25 (one-flesh next to naked — the marriage-document hint) · va-yevie-ha letter-sum 24 = the 24 bridal adornments (Kitzur/Zohar via Minchat Shai; gematria-class — candidate for a future verifier check type).
DISCREPANCIES
None.
INSTRUMENT GAPS (filed)
- G09-10 the paseq: Minchat Shai reads a paseq (separator bar) between Elohim ("God") and tardemah ("deep sleep") "for the honor of the Name" — SNAPSHOT encodes ZERO paseq marks corpus-wide (verified this audit): a newly discovered instrument gap, affects all future accent-claims of this class.
- G09-11 ha-tzela pointing dispute (manuscript variants not encoded).
TIMING + PROCESS NOTES
Elapsed for the oral layer: ~22 minutes (span query + local reads + manifest + verifier + this record), vs ~45 min for gen_08's web-fetched layer. Zero fetch-summary garbles — verbatim local text end to end. Notes: (1) paseq gap discovered — record once, stop re-checking; (2) gematria-class claims (letter sums) are mechanical — worth a verifier check type if they recur; (3) Bavli Hebrew is local (Data/bavli_*.json) but NOT in export_texts — indexing it would make Talmud loci readable in-line too.