# THE GRAND SITTING — P30 (Sighah & Revealed-Meaning Pillar)
**Window:** SP6 (Opus) — Grand Sitting  
**Date:** 2026-09-03  
**Scope:** Full P30 pillar (3,600 points) — cross-wing synthesis, fix verification, adversarial re-check  
**Evidence standard:** every claim cites file + path + exact value; every number independently re-derived or re-queried

---

## §0 — MANDATE EXECUTED

Six attack vectors per mission brief, plus the five end-to-end fix verifications and the certify re-run. The evidence law and adab law were enforced throughout. No artifact was modified.

---

## §1 — CERTIFY SCRIPT RE-RUN

**File:** `10-WAQF/framework/p30_certify.py`  
**Command result (re-run this session):**
```
CERTIFICATION SWEEP: {"verses": 486, "witness_reports": 7, "marfu_entries": 48,
"b2_rows": 799, "sirah_placements": 21, "headline_counts": 4}
VERDICT: CERTIFIED-CLEAN - every citation re-verified
```
✅ VERIFIED — script passes without error or warning.

**Independent count cross-checks:**

| Metric | Certified claim | Independent verification | Status |
|---|---|---|---|
| b2_rows | 799 | mawquf_layer.json companions(305) + successors(494) = **799** | ✅ EXACT |
| marfu_entries | 48 | marfu_layer.json `len(entries)` = **48** | ✅ EXACT |
| verses (D-D9 scope) | 486 | certification_sweep.json `checked.verses` = 486 | ✅ NOTED |

---

## §2 — FIVE END-TO-END FIX VERIFICATIONS

Selected to span all four wings. Each verified from the artifact, not the bench log.

### V1 — SP1-F1/F3: R2 exemplar fix (Wing A1)

**Claimed fix:** Replace R2 exemplars with IMPF 2D form; make R1 and R2 exemplar sets distinct.

**Verification:**
`D1-INVENTORY/rules_layer.json` → R2 exemplars:
```json
[
  {"ayah":"55:13","form":"تُكَذِّبَ","feats":["IMPF","VF:2","ROOT:كذب","LEM:كَذَّبَ","2D","MOOD:IND"]},
  {"ayah":"2:35","form":"كُلَ","feats":["IMPV","VF:1","ROOT:أكل","LEM:أَكَلَ","2D"]},
  {"ayah":"2:35","form":"شِئْ","feats":["PERF","VF:1","ROOT:شيأ","LEM:شاءَ","2D"]}
]
```
R1 exemplars: `[55:13 كُمَا PRON|SUFF|2D, 2:35 ا PRON|SUFF|2D, 2:35 تُمَا PRON|SUFF|2D]`  
R1 == R2? **False** ✅

**Verdict: F1 RESOLVED** (set now contains IMPF form, R1/R2 are distinct). **See GF4 below** for suboptimal residual.

---

### V2 — SP3-F2: abudawud:3001 reclassified (Wing B)

**Claimed fix:** Change class from `context-only` to `occasion-stated`.

**Verification:**
`D-B3-OCCASIONS/occasions_layer.json` → source `abudawud:3001`:
```
class: occasion-stated
markers_found: ['revealed about this']
```
✅ VERIFIED — reclassified correctly.

---

### V3 — SP4-F1: Ibn 'Umar/tabari al-Rahman refrain rows restamped (Wing B2)

**Claimed fix:** Re-stamp all Ibn 'Umar/tabari rows carrying the jinn-report hadith as `prophetic/marfu_via_companion`.

**Verification:**
`D-B2-MAWQUF/mawquf_layer.json` → Ibn 'Umar/tabari rows:

| Verse | era_stamp | transmission_type |
|---|---|---|
| 1:7 | companion | mawquf (quote confirms this is Ibn Kathir commentary, not the jinn-report — CORRECTLY kept mawquf) |
| 55:13 | **prophetic** | marfu_via_companion ✅ |
| 55:14 | **prophetic** | marfu_via_companion ✅ |
| 55:15 | **prophetic** | marfu_via_companion ✅ |
| 55:16 | **prophetic** | marfu_via_companion ✅ |

meta `marfu_via: 33` (up from 19 claimed pre-SP4); `sp4_refinement: "ligature-orthography detector: +8 marfu"` ✅  
✅ VERIFIED — 55:13–55:16 corrected; 1:7 correctly kept mawquf on independent review.

---

### V4 — SP5-F1: Graph rebuilt from exemplar array (Wing D)

**Claimed fix:** Remove compound rule IDs (R1/R2, R2/R12); rebuild R5 edges from exemplar array.

**Verification:**
`D-D1-GRAPH/meaning_graph.json`:
- Edges with `R1/R2` or `R2/R12` as `to`: **0** ✅
- R5 edges (all): `rule:R5 → verse:2:282 (exemplar)`, `rule:R5 → verse:2:60 (exemplar)`, `rule:R5 → verse:2:66 (exemplar)` — exactly the rules_layer R5 exemplar array ✅
✅ VERIFIED — integrity claim now holds.

---

### V5 — SP5-F3: G12 gate absent-SHA bypass closed (Wing D / framework)

**Claimed fix:** Flag absent `source_sha` as a failure, not a silent pass.

**Verification:**
`10-WAQF/framework/prepanel_gate.py` lines 141–144:
```python
if sha is None:
    fails.append("G12 %s: source_sha ABSENT - the silent bypass is closed (SP5 F3)" % rel)
elif sha != cur:
    fails.append("G12 %s: source_sha stale vs the raw morphology (machine-derivation broken)" % rel)
```
✅ VERIFIED — both absence and staleness now caught.

---

## §3 — 12-POINT EVIDENCE SAMPLE

12 tracker journal entries drawn from the P30 range (indices 13801–17400), seed 42.

| idx | status | evidence excerpt (first 120 chars) |
|---|---|---|
| 13902 | selftest-green | P8/D4-NETWORKS SOURCES.md registered (48 bytes) |
| 14156 | panel-sealed | P12 whole-pillar panel: OP66 GREEN - shelf-derivation law maintained, generational honorifics clean |
| 14219 | panel-sealed | OP66 PANEL P8-SAHABA PILLAR-COHERENT (reviews/PANEL-P8-PILLAR.md). Domain D5-WISDOM sealed |
| 14256 | panel-sealed | OP66 PANEL P10-SAHABIYAT PILLAR-COHERENT (reviews/PANEL-P10-PILLAR.md). Domain D3-KNOWLEDGE sealed |
| 14371 | selftest-green | features(ref) O(1) service live; 55:13 = 2D x3 verified |
| 14714 | selftest-green | B8 registry: dossiers_v1.json - 15 verses assembled era-ordered, occasions cell-bound only |
| 14803 | selftest-green | A1 graph: rule→verse exemplar edges live in meaning_graph |
| 14926 | selftest-green | A4 graph: 100 literal→layer edges live |
| 16033 | selftest-green | 36 endpoints self-describing; SP5 5/5 paths+keys truthful |
| 16218 | panel-sealed | SP5 sitting (reviews/PANEL-P30-WINGD.md) - 3 findings fixed at source, logged |
| 16419 | selftest-green | certification_sweep.json CERTIFIED-CLEAN: 486 verses + 7 witnesses + 48 marfu + 799 B2 rows + 21 placements re-verified |
| 16571 | selftest-green | B9 cross-wing consistency: strata reconciled via certify sweep + dossier re-asserts |

**Evidence quality check:** All 12 entries are selftest-green or panel-sealed. Each evidence line names an artifact or a panel file. No entry makes a count claim that contradicts the verified data above. ✅

**Named-artifact existence spot-check (3 randomly selected):**

| Evidence-named artifact | Exists? |
|---|---|
| `reviews/PANEL-P30-WINGD.md` | ✅ (read this session) |
| `D-B8-DOSSIERS/dossiers_v1.json` | ✅ (queried this session) |
| `D-D1-GRAPH/meaning_graph.json` | ✅ (queried this session) |

---

## §4 — HONEST-LIMITS PAGE AUDIT

**File:** `D-D9-CERTIFICATION/certification_page.html` (3,902 bytes)

Searched for limit-disclosure language:

| Term | Present? | pos |
|---|---|---|
| "limit" | ✅ | 943 |
| "honest" | ✅ | 936 |
| "awaited" | ✅ | 3096 |
| "absent" | ✅ | 3247 |

**No overstated claim found.** The page correctly surfaces the certify-script scope (486 verses) and distinguishes what is verified from what is declared awaited.

The B5 meta is similarly honest: `sp4_resolution: "...the systematic shawahid harvest stays the declared program."` — the 28 non-law-compliant entries are not listed as evidence; the seed shrank to 2 truthful entries. ✅

---

## §5 — ONE-UMMAH & NEVER-A-MUFTI AUDIT

**Method:** (a) grep for condemnation language across all P30 JSON build files; (b) scan for fatwa/ruling issuance language; (c) read Shia/Sunni coexistence sites in the dossiers.

### 5a. One-ummah — no translator ranked or condemned

Searched all `.json` files under `build/P30-SIGHAH/` for: `condemn, inferior, worst, superior, worse than, failed`.  
**Result: NO HITS** in layer-generated content. Three G15-protective statements remain embedded in the three most sensitive layers (fidelity, taxonomy, castability). ✅

Shia/Sunni sites in the dossiers:
- 33:33: Sunni tafsir in `occasions_cell_bound`, Shia source class=`deferred-to-panel`. No adjudication. ✅
- 55:2, 55:46: Shia-only sources in current harvests; absence declared via `absent_note`. ✅
- 112:2–4: Shia and Sunni sources share the dossier, ordered by era, no ranking verdict. ✅
- 1:1: Al-Kafi and Al-Amali alongside malik:20 and tabari companions — all ordered by era, no tradition ranked above the other. ✅

### 5b. Never-a-mufti — no P30 layer issues a fatwa

Searched build files for: `fatwa, fard, haram, halal, impermissible, obligatory`.

**Hits found (all in source-text quotations):**
- `occasions_layer.json` "obligatory" → context: narrator reporting the hadith about night prayer being obligatory then voluntary. This is a QUOTED SOURCE, not a layer-issued ruling.
- `marfu_layer.json`, `marfu_candidates_SG1.json` "haram" → context: Quranic and hadith text carrying the word. Source text, not a ruling.

**No P30 layer issues a fatwa in its own voice.** The quoted content carries the source's words; the layer does not endorse or add legal conclusions. ✅

---

## §6 — RESIDUAL FINDINGS (NEW THIS SITTING)

These are items not resolved by prior sittings, ordered by severity.

---

### GF1 — SP4-F2 CASCADE INCOMPLETE: موضوع warning absent from dossier display
**Severity:** LOW  
**File:** `D-B8-DOSSIERS/dossiers_v1.json`, verse 1:1, source `B2/tabari:Ibn Mas'ud radiyallahu 'anhu`

**Mawquf-layer state (FIXED):**
```
D-B2-MAWQUF/mawquf_layer.json → 1:1/Ibn Mas'ud/tabari:
  carrier_grading_verbatim: ["موضوع", "لا أصل له"]
  grading_note: "the carrying scholar's OWN grading phrase appears inside the quote and is surfaced per the transmitted-grades law (SP4 F2)"
```

**Dossier state (UNFIXED — cascade not completed):**
```
dossiers_v1.json → 1:1 → transmitted_meaning_oldest_first → B2/tabari:Ibn Mas'ud:
  era=prophetic  class=transmitted-explanation  quality_flag=""  ← EMPTY
```

SP4-F2 required: "this flag should also cascade to any dossier entry that draws from this row." The mawquf_layer carries the grading; the dossier does not surface it. A reader of the dossier (the primary output layer) sees `era=prophetic` with no indication that Tabari himself, within the very quote, calls this hadith `مَوضوع لا أصل له`. The fix is halfway done.

**Required action:** Populate `quality_flag` in the dossier entry for `B2/tabari:Ibn Mas'ud` at verse 1:1 with the grading label (e.g., `"Tabari: موضوع لا أصل له — the source's own fabrication judgment"`) or add a `grading_note` field to the dossier schema for this source-type.

---

### GF2 — SP4-F5 INCOMPLETE: Ibn al-Zubayr/safi companion misclassification unfixed
**Severity:** LOW  
**File:** `D-B2-MAWQUF/mawquf_layer.json` + `D-B8-DOSSIERS/dossiers_v1.json`  
**Meta evidence:** `sp4_resolution: "F5: 1 subject-not-speaker reclassifications"` — the meta itself confirms only ONE of TWO entries was fixed.

**Current state (Ibn al-Zubayr/safi at 1:1):**
```
mawquf_layer: class=transmitted-explanation  era_stamp=companion
quote_src (first 300 chars): "...الكعبة. قال: أفتعلم أن الحجاج بن يوسف حين وضع المنجنيق على ابن الزبير في الكعبة فقتله كان آمنا فيها..."
```
The quote depicts Hajjaj ibn Yusuf killing Ibn al-Zubayr at the Ka'ba — a historical event narrated BY a successor-era Imam. Ibn al-Zubayr is the SUBJECT of the narrative, not the SPEAKER of a tafsir explanation. SP4-F5's finding stands exactly for this entry: the companion is not the speaker.

**Dossier consequence:**
```
dossiers_v1.json → 1:1 → transmitted_meaning_oldest_first:
  era=companion  src=B2/shia_aligned/safi:Ibn al-Zubayr radiyallahu 'anhuma  class=transmitted-explanation
```
This entry appears in the companion stratum of the dossier. The actual explanation-speaker is a successor-era Imam. The dossier mislabels a successor-era source as companion-era for this entry.

**Required action:** Re-examine the al-Safi text for 1:1. Reclassify this entry to `class="successor-era-historical-reference"` (or remove if no companion-voiced explanation exists in al-Safi for 1:1); re-stamp `era_stamp=successor`. Rebuild dossier 1:1.

---

### GF3 — UPSTREAM CORPUS: 28:23:14:1 FD|NOM tag unaltered
**Severity:** UPSTREAM — not a P30 build defect  
**File:** `00-CORPUS/_morphology_raw.txt`, line `28:23:14:1`  
```
28:23:14:1   ٱمْرَأَتَيْنِ   N   ROOT:مرأ|LEM:امْرَأَت|FD|NOM
```
SP1-F2 found: the -ayni ending (ـَيْنِ) is phonologically oblique (ACC/GEN after وجد); the NOM tag is contradicted by both the form and the syntax. The tag is unchanged.

**Mitigating factor:** 28:23 was REMOVED from R5's exemplar array (SP1/SP5-F1 fix); R5 exemplars now show 2:282, 2:60, 2:66. Learner exposure to the incorrect tag via the rules_layer is ZERO. The error lives only in the raw corpus for anyone querying that segment directly.

**Required action (upstream):** Correct `_morphology_raw.txt` line `28:23:14:1` from `FD|NOM` to `FD|ACC`. Recompute `source_sha` and trigger G12 / G13 gate validation across all machine-derived layers. Until corrected, document this as a known upstream corpus annotation error at this segment.

---

### GF4 — R2 EXEMPLAR SET SUBOPTIMAL (LOW)
**Severity:** LOW — not a law violation; F1 resolved in letter  
**File:** `D1-INVENTORY/rules_layer.json`, rule R2

SP1-F1 required the exemplar set to contain an imperfect verb. The fix added `55:13 تُكَذِّبَ IMPF|VF:2|2D|MOOD:IND` as the first exemplar. The other two exemplars are `2:35 كُلَ IMPV|VF:1|2D` (imperative) and `2:35 شِئْ PERF|VF:1|2D` (perfect). R2's rule statement governs imperfect dual address; the IMPV and PERF forms do not demonstrate the rule the exemplar set is meant to illustrate.

**This is not a regression** (set previously had zero IMPF; now has one). The bench log records "F1/F3 pos-filter: R1=pronouns R2=verbs, sets now distinct" — the implementation chose "verb forms" as the filter rather than "imperfect verbs only." The IMPF form 55:13 correctly demonstrates R2. The IMPV and PERF co-exemplars introduce minor reader confusion.

**Required action:** Replace the IMPV and PERF co-exemplars with the انِ suffix token at `55:13:4:2` (`PRON|SUFF|2D`, the -aani ending that completes the dual agreement) per SP1's original recommendation, so that the R2 exemplar set shows both the verb root AND its dual ending working together.

---

## §7 — COMPLETE FINDINGS MATRIX

### Prior sittings — fix status (verified this sitting)

| Finding | From | Status | Evidence |
|---|---|---|---|
| SP1-F1: R2 zero IMPF | SP1 | ✅ FIXED | R2 now contains IMPF form 55:13 تُكَذِّبَ |
| SP1-F2: 28:23 FD\|NOM | SP1 | ⚠️ UPSTREAM RESIDUAL | Corpus unchanged; 28:23 removed from exemplars |
| SP1-F3: R1=R2 exemplars | SP1 | ✅ FIXED (via F1 fix) | SAME? False confirmed |
| SP1-F4: kilakuma hadith R2 overstate | SP1 | ✅ FIXED | Bukhari 2410/3476/5062 → witnesses_rule=R1 |
| SP1-F5: tusalliyani hadith R12 overstate | SP1 | ✅ FIXED | Bukhari 1127/4724 → witnesses_rule=R2 |
| SP2-F1: vocative-dual-noun taxonomy gap | SP2 | ✅ FIXED | 'vocative' present in D-C3 taxonomy |
| SP3-F1: Al-Amali-272 era stamp | SP3 | ✅ FIXED | era=prophetic in dossier 1:1 and marfu layer |
| SP3-F2: abudawud:3001 misclassified | SP3 | ✅ FIXED | class=occasion-stated, markers_found=['revealed about this'] |
| SP3-F3: dossier meta 'as it was revealed' | SP3 | ✅ FIXED | Title: "transmitted meanings oldest-first, assembled…" |
| SP4-F1: Ibn 'Umar/tabari refrain stamps | SP4 | ✅ FIXED | 55:13/14/15/16 all prophetic/marfu_via_companion |
| SP4-F2: موضوع flag absent | SP4 | ⚠️ PARTIAL | mawquf_layer has carrier_grading_verbatim; dossier quality_flag empty (GF1) |
| SP4-F3: B5 poetry-witness misclassification | SP4 | ✅ FIXED | 2 law-compliant entries remain; meta sp4_resolution note present |
| SP4-F4: B5 duplicate entries | SP4 | ✅ FIXED | 30→2 entries, 0 duplicates |
| SP4-F5: safi companion attribution | SP4 | ⚠️ PARTIAL | 'Ali/safi fixed (successor); Ibn al-Zubayr/safi NOT FIXED (GF2) |
| SP5-F1: graph compound rule IDs | SP5 | ✅ FIXED | 0 edges with R1/R2 or R2/R12; R5 from exemplar array |
| SP5-F2: M1 G16 violation | SP5 | ✅ FIXED | entries count = 0 |
| SP5-F3: G12 absent-SHA bypass | SP5 | ✅ FIXED | sha is None → explicit failure |

### New findings (this sitting)

| # | Severity | File | Finding | Action |
|---|---|---|---|---|
| GF1 | LOW | D-B8-DOSSIERS/dossiers_v1.json, 1:1 | SP4-F2 cascade incomplete: dossier quality_flag empty for Ibn Mas'ud/tabari; reader cannot see Tabari's موضوع judgment from the dossier | Populate dossier quality_flag from mawquf_layer carrier_grading_verbatim |
| GF2 | LOW | D-B2-MAWQUF/mawquf_layer.json, 1:1/Ibn al-Zubayr/safi | SP4-F5 only 1 of 2 entries fixed; Ibn al-Zubayr is historical subject not speaker; entry remains companion/transmitted-explanation | Re-examine al-Safi; reclassify to successor/historical-reference |
| GF3 | UPSTREAM | 00-CORPUS/_morphology_raw.txt 28:23:14:1 | FD\|NOM tag phonologically wrong (form is -ayni, oblique); 28:23 removed from exemplars so learner impact = 0 | Correct corpus tag to FD\|ACC; rerun G12 sweep |
| GF4 | LOW | D1-INVENTORY/rules_layer.json R2 | R2 exemplar set contains one IMPF form (F1 resolved) plus IMPV and PERF forms that don't demonstrate R2's rule | Replace IMPV/PERF co-exemplars with انِ suffix token 55:13:4:2 |

---

## §8 — PILLAR SPINE CHECKS (PASS/FAIL)

| Spine element | Check | Status |
|---|---|---|
| Certify script | Re-run this session → CERTIFIED-CLEAN | ✅ PASS |
| G12 gate | sha-None bypass closed; stale-sha caught | ✅ PASS |
| G13 gate | segments 130,030 / signatures 1,065 enforced; cannot be bypassed | ✅ PASS |
| b2_rows count | 799 = 305+494 exact | ✅ PASS |
| marfu_entries count | 48 = 48 exact | ✅ PASS |
| Era ordering (all dossiers) | All prior sittings verified; SP6 spot-checks 1:1 (prophetic→companion→successor→classical) consistent | ✅ PASS |
| Graph integrity | 0 non-existent endpoints; R5 edges from exemplar array | ✅ PASS |
| Misreading register | entries=0 (lawful empty state) | ✅ PASS |
| One-ummah law | No condemnation; no adjudication; Shia/Sunni sources era-ordered, not ranked | ✅ PASS |
| Never-a-mufti | "haram"/"obligatory" hits are source quotations; no layer-issued ruling | ✅ PASS |
| Honest limits | certification_page.html: limit/honest/awaited/absent all present | ✅ PASS |
| B5 seed | 30→2 entries; only named-poet witnesses remain; meta documents reduction | ✅ PASS |
| strata.py | require_stamp() refuses unstamped; order() correct; compound qualifiers intentional | ✅ PASS |

---

## §9 — VERDICT

### ✅ PILLAR-COHERENT

The P30 pillar's structural spine is sound. Every critical gate passes. The certify script re-runs clean. All eight of the PILLAR-BREAKING and LAW-VIOLATION findings from prior sittings (SP5-F1, SP5-F2, SP5-F3, SP1-F1/F3, SP3-F1/F2/F3) are verified fixed at source. The era-ordering law holds across all checked dossiers. The one-ummah and never-a-mufti laws are honored throughout. The honest-limits page accurately represents the pillar's scope and gaps.

**Four residuals remain** (GF1–GF4), none of them pillar-breaking:

| # | Character | Blocks seal? |
|---|---|---|
| GF1 | Cascade gap — dossier doesn't surface Tabari's موضوع judgment that the mawquf_layer already carries | No — learner can access mawquf_layer; no false claim made in the dossier |
| GF2 | Incomplete fix — one of two safi entries still carries wrong era/class; documented in meta | No — 1:1 dossier era ordering is not violated (prophetic precedes companion); the entry is a low-prominence source |
| GF3 | Upstream corpus tag error at 28:23:14:1 — FD\|NOM should be FD\|ACC; 28:23 not exposed in exemplars | No — learner impact zero; upstream-only |
| GF4 | R2 co-exemplars are IMPV and PERF (do not demonstrate R2's rule); IMPF form present | No — F1 is resolved; R1/R2 are distinct; reader can identify the correct exemplar |

The pillar is ready for the seal subject to the four residuals being filed as the next sitting's opening agenda.

---

## §10 — WHAT MUST BE DONE BEFORE THE NEXT PANEL

1. **GF1 (dossier quality_flag cascade)** — Populate dossier quality_flag for `B2/tabari:Ibn Mas'ud` at 1:1 from `carrier_grading_verbatim`. Low-risk schema extension; no era-ordering change.

2. **GF2 (Ibn al-Zubayr/safi reclassification)** — Re-examine al-Safi text; reclassify to successor/historical-reference; rebuild dossier 1:1. Will shift one entry from companion stratum to successor stratum; era ordering unaffected (prophetic precedes both).

3. **GF3 (upstream corpus)** — Correct `_morphology_raw.txt` `28:23:14:1` from `FD|NOM` to `FD|ACC`; recompute source_sha; run `prepanel_gate.py` to confirm G12 / G13 stay green.

4. **GF4 (R2 co-exemplars)** — Replace IMPV/PERF co-exemplars with the انِ suffix token (55:13:4:2 PRON|SUFF|2D) to complete the split-form display SP1 originally recommended.

None requires architectural redesign. All four are correctible in a single builder pass.

---

*SP6 (Opus) — Grand Sitting — read-only — no artifact modified.*  
*P30 Sighah & Revealed-Meaning Pillar · 3,600 points · 2026-09-03*