Qamus particle family P-03 — 16 entries (question/response + operator particles)

[theislampill]

Qamus particle family P-03

Generated from repo ledgers at main@3396d79 (2026-08-05). This issue is a public mirror of the committed ledgers; on any conflict the ledgers win. Evidence in comments must cite repo paths, PR numbers, or blocker ids — never private-workspace packet contents or server topology.

What this family is

Particle work-ordering family P-03: 16 entries — أَ, سَـ, الْ, يَا, سَوْفَ, كَأَيِّن, رُبَّمَا, كَيْفَ, أَيّ, أَنَّى, كَمْ, أَيْنَ, مَتَى, أَيَّانَ …
(pages: p001, p004, p010, p020, p028, p031, p032, p042, p043, p044, p046, p047, p049, p050, p053, p055; source: qamus/data/particle-tranche-membership.json). Function families: interrogative hamza (istifhām/taswiya); future tanfīs sa-; definite article al-; vocative yā; future tanfīs sawfa; kam-khabariyya analogue ka-ayyin; taqlīl/takthīr rubbamā; interrogative kayfa; interrogative/relative ayy; interrogative annā; interrogative/khabariyya kam; interrogative ayna; interrogative matā; interrogative ayyāna; response balā; response ī (oath-assent).

Plan renumber (owner decision 2026-08-05): the original 49-entry P-03 long tail was renumbered into three flat tranches cut on shared analytical machinery — P-03 question/response + operator particles (16), P-04 demonstratives + relatives (18), P-05 personal pronouns (15) — recorded in qamus/data/particle-tranche-membership.json. The particle plan is P-00..P-05 at 12/17/22/16/18/15.

The denominator that makes this the top-yield lane: the particle universe is ~50,263 candidate
occurrences with ~109,364 appearances, and 99.1% of those appearances sit on V/N pages as
example-āyah context words — currently certified: none for every row. The particle pages' own
spans are 983/984 rich live, but that old-era richness does not transclude; only certification does.

Impact on transclusion & the sarf/naḥw lattice projection

This is the purest transclusion lane in the programme: certify one particle occurrence and the
projection-hash parity invariant renders it rich on every card it appears on — across all 24 VN
tranches — with zero per-page authoring (measured on the VN-00 readback pilot: ~2,690 rendered-span
edges corpus-wide from one window's accounting). Surface match never authorizes reuse (og-1/og-2):
membership goes through the candidate matrix → disposition → certification rungs, never lookalike copying.
When this family closes, post the measured unlock counts on the affected VN issues.

Impact on the flywheel & repo

Family certification exercises the sarf/naḥw skills on real corpus at scale; every abstention,
disagreement, or defect (e.g. clitic-host segmentation, allomorph families like لَهُمْ/لَكُمْ) returns
to /fusha/ as fixtures, misconception rows, or registry keys. Dual-function and homograph rows
exercise rival preservation — outcomes feed the naḥw skill's attributed-alternatives machinery.

Closure conditions (ladder)

• Family pre-flight verified per owner ruling: entry ids, sense inventories, candidate populations against the example-āyah universe, p007-machinery reuse measured and reported
• Every candidate occurrence dispositioned: certified (direct-source or two-vote rung of tools/certify_typed_fact.py), guarded projection, or explicit pending/scholar state with cause
• Projection edges emitted to all V/N-page appearances with hash parity; per-VN-tranche unlock counts posted on the affected VN issues
• 🔒 Deployed live (owner-gated): deploy window updates qamus.dawah.wiki; family occurrences render rich on particle pages AND on V/N cards — cite the deploy receipt
• Flywheel return: fixtures/misconceptions/registry rows emitted to /fusha/, with paths

Update & closure discipline

• The executor updates this issue as work lands: check boxes with a one-line evidence citation
  (PR number + ledger path), and post a comment when another issue's closure changes this window's
  measured state (e.g. a particle-family closure lighting up context words here — cite the count).
• This issue CLOSES only when BOTH are true: the repo ledgers record the closed state AND the live
  website renders it. The deploy rung is owner-gated (docs/blockers.yaml → owner-deploy-packet);
  everything above it is executor work. Candidate-plane completion with no deploy = issue stays open.
• The closing comment must state the realized impact: occurrences accounted, facts certified, cards
  now rich live, cross-page appearances lit elsewhere (name the issues), and what the flywheel
  returned to /fusha/ (fixtures, misconception rows, drill keys — cite paths).