FESTIVAL_SLICE_EXT_CH08_09.md

world/regions/ethiopia/page/FESTIVAL_SLICE_EXT_CH08_09.md

Festival slice extension — CH_08 Swahili Coast and CH_09 Ethiopia

Proposal-tier. Seven proposed T0_Festival_Registry rows in the landed FST_0001-0016 register, authored

against the canon nodes for CH_08 and CH_09 and their region pages. The sixteen landed rows are

READ-ONLY and are not edited by this lane. No festival_id is minted here — every row carries

(DIRECTOR ALLOCATES) and the director allocates on apply. Windows are SHAPES per LW-3; participation is

opt-in and non-gating throughout. Care authority is CVD §17.1, and both chapters carry living religious

traditions, so ceremony sits at reference register and no protected liturgy or ceremonial text is

reproduced anywhere below.

Derivation

Work kind resolved live with python harness/route.py registry --chapter CH_08 --region swahili_coast.

The read-set it printed was honored in order (CLAUDE.md authority hierarchy; the spine node; the row-struct

spec; the extension, FK and generation-lifecycle manifests). The registry's owning T1 doc is the gap the

pipeline ledger already records — B7 has NO T1 owner, and the charter is the DOC_MAP LW-5 row with care

authority at CVD §6 and §17.1 — so the owning authority read in its place was the CVD §17.1 Care Doctrine

plus the two region pages and the two spine nodes.

DERIVED FROM:

NOT DERIVED (authored judgment, and why it had no canon home):

The proposed rows — CH_08 Swahili Coast

Row A — the New Year at Makunduchi

Row B — the mizimu dues across the millet cycle

Row C — the safe-voyage offering and the votive ship engraving

The proposed rows — CH_09 Ethiopia

Row D — the tabot procession

Row E — the four annual liturgies upon the ark-object

Row F — Mäsqäl and Qeddus Yohannes

Row G — the three-pour buna ceremony

Named gaps

NAMED GAP 1 — CH_08 has no attested festival in its own present frame

The swahili_coast shelf attests exactly ONE festival, and it belongs to Makunduchi on southern Unguja and

to the modern window. The chapter actually plays in the 1290-1400 Mahdali present frame at Kilwa, and for

that frame the shelf attests no calendar festival at all. Rows A and B are therefore both Section-18

modern-window rows and Row C is the only proposed row reachable in the present frame. This is a substrate

gap and not an authoring one: the honest fix is the Section-16 MUST-ACQUIRE (Horton and Middleton for the

Section-4 depth; the Mackenrodt and Giles acquisition that already owes the waganga and ngoma verification),

not a minted medieval Kilwa festival. A festival invented onto the present frame would be uncited by

construction and would read as authoritative cultural detail while resting on nothing.

NAMED GAP 2 — no Mijikenda rite is authored, and none should be yet

The Mijikenda carry a whole beat (CH08_B07), a protected UNESCO-inscribed sacred grove, and one of the

chapter's two positive recognitions — and the region page records, positive-controlled, that this shelf

attests NOTHING about them. No Mijikenda festival, rite or observance is proposed here. A row authored onto

a living people's protected sacred ground from zero substrate is precisely the §17.1 authenticity failure

the page's own naming block guards against. The row is owed to the MUST-ACQUIRE Mijikenda source and to

nothing else.

NAMED GAP 3 — the weekly window has no window_kind

Ethiopia attests two weekly fast days (Wednesday and Friday) and a Saturday-and-Sunday double observance

whose two names carry the confusion in their own etymologies, and Section 18 makes both mechanically live

(the fast days change what a household will feed the player). Both are rite-shaped and both are cited. The

register's window_kind vocabulary carries calendar_annual, event_triggered, pawukon_cycle, lunar and daily,

and none of them is a weekly cycle. No kind is minted here: a new vocabulary value is a registry-wide

decision the director owns, not one a slice-extension lane should take inside a row. If the director rules

a weekly kind in, the Ethiopian fast-and-Sabbath cycle is its first and strongest member and this lane can

author the row in one pass.

NAMED GAP 4 — the Sabbath argument is a window but not a festival

The region page's own festival lane carries five bullets and this proposal realizes four of them. The

fifth, the Sabbath question, is a live and locally contested matter of practice staged as an argument the

player can walk into in the Section-18 Portuguese-contact window, with the parish never humiliated and

neither side written as a villain. It has a window (the contact stratum) and a site (the church), but it

is a dispute rather than a rite, so it does not take a Festival_Registry row. It is recorded here so a

later lane does not read the four-of-five coverage as an oversight.

Findings for the director

FINDING 1 — ZONE GAP, CH_08 Unguja

The CH_08 Zone Catalog carries five zones, all Kilwa-bay or Kenyan-coast ground; Zanzibar appears only as an

inter-chapter harbour waypoint in the spine's travel line. Rows A and B are both Makunduchi rites on

southern Unguja and are therefore held at region_level with the gap declared in-row, per the FST_0005 and

FST_0006 precedent. If the build wants the New Year staged rather than referenced, it needs an Unguja zone

first — and that zone would sit under the same elevated care_tier logic the page applies to Pemba (a real

named living island community).

FINDING 2 — the naming discipline is load-bearing and a later lane will be tempted to break it

The swahili_coast page runs a literal term census across all 472 chunks of its region-filtered shelf, and

Row A's festival is named only as the page names it. A widely-known local name for a Zanzibari New Year

festival exists in general knowledge and is NOT on this shelf; it is deliberately absent from the row, and

a later pass that adds it without an acquisition behind it would be supplying uncited cultural detail that

flows straight into the Section-17 localization glossary. Same discipline, same reason, for ngoma, waganga,

jini and majini: none appears in any cell above.

FINDING 3 — care adjudication, the stick contest kept

Row A offers the player a role in the young men's stick contest between the town's halves rather than

routing around it or reframing it as a dance. Reasoning: the source records the fight plainly as part of

the day; CVD §17.1's ruled calibration holds that care-tier inflation is itself a defect and that thinning

grounded material to the inoffensive is not care; and FST_0014's landed precedent keeps a real buffalo

sacrifice with its killing and prestige intact for exactly this reason. Participation stays opt-in and

non-gating, which is where the player-side protection lives.

FINDING 4 — care adjudication, the modern-window typing of both CH_08 calendar rows

Rows A and B are typed to the Section-18 modern window rather than promoted into the medieval present.

This is both a chronology call (the fieldwork is contemporary and the source traces the festival's

twentieth-century reshaping) and a care call: presenting a living community's current practice as its

fourteenth-century self would flatten four hundred years of that community's own history into set-dressing.

The cost is NAMED GAP 1, and the gap is the more honest outcome.

FINDING 5 — care adjudication, the assist ceiling at Row E

Row E is the only row in this extension whose participation ceiling is a care boundary rather than a player

choice. The player may walk the route and carry the canopy pole; the player may never take a bearer's

place, never handle the object, and the chamber is never entered. That framing is the region page's own,

and it is grounded in the tradition's own reasoning about the ark rather than imposed from outside. The

assist value carries it exactly (the FST_0006 precedent), which is why the row does not use

opt-in-participate.

FINDING 6 — the zema and the SQ7 case are deliberately not pre-empted

No Ethiopian Orthodox liturgical text, chant, chant-mode or notation is authored anywhere in these rows.

Saint Yared's zema is Thread 21's anchor and the region page records an unusually strong SQ7 Sorcery/Music

Theory case for this region — and then declines to stage it, because the spine assigns SQ7 no stage at Ch 9

and the page holds that any SQ7 seed here is a Josh call at the range review. This lane holds the same line:

Row D carries the procession's sound as attested instruments and public ambience, and nothing more.

FINDING 7 — Rows D and E are two rows on purpose

Both concern ark-object carriage, and merging them would look tidier. They are kept separate because they

differ on four of the thirteen columns in ways that matter: window (annual at Timqat plus many occasions

versus four fixed liturgies), site (region-wide versus the Aksum royal precinct), participation

(opt-in-participate versus assist), and care ceiling (nothing consecrated reproduced versus the inviolable

chamber and the never-handled object). The landed register already carries three Balinese rows of closely

related devotional grammar (FST_0008, FST_0011, FST_0012) on the same logic. If the director merges them,

the assist ceiling is the half that must survive.

FINDING 8 — no canon conflict found, with one standing declaration worth the director's eye

Nothing in these rows required a change to a spine node, a region page, a T1 doc or a registry row, and

none was made. One pre-existing declaration is repeated here so the festival lane does not re-derive it:

the CH_08 spine's asset anchors name the ngoma ya majini possession-rite drumming, and swahili_coast.md's

Section-5 naming block records ngoma at zero occurrences across the processed shelf, with the term consumed

as spine canon and routed to the Section-16 acquisition. That is already resolved on the page as a

consume-not-re-author case. No festival row above uses the term, and none should until Mackenrodt and Giles

land.

FINDING 9 — row-count note

Three rows for CH_08 and four for CH_09, against the lane's roughly-one-to-three expectation. CH_09's fourth

row is Row G, the daily buna baseline, and it is included on the FST_0010 precedent and because its real

function is prophylactic: it fixes the chapter's most-played domestic scene as reference-register and

unauthored-beyond-three-names before a build lane invents procedure for it. If the director wants the count

held to three, Row G is the one to drop, and dropping it should carry the care_note's unauthored-procedure

constraint forward to whichever artifact stages CH09_B01.

Generated by harness/site/structure_site.py — the URL path is the repo path. review root