swahili_coast.md

canon/tiers/02_Tier_2_Region_Pages/swahili_coast.md

Region Page — Swahili Coast

Canonical region page (T1_Region_Page_Template [ACTIVE v2.0]), cut to the flores_island pilot's

section architecture. It consumes the v3.1 spine's typed layer for CH_08 (the join root) and authors

only the region-scope substrate the spine leaves downstream. The region's identity resolves in

T0_Region_Index; the page passes harness/check_region_page.py. Reveal discipline binds at page grain

exactly as at spine grain: per Josh's ruled Q1 handling no House name, family, or pattern reaches any

player surface at Ch 8 — the player receives a hostile recognition and an unformed unease and nothing

more — and the Grand Sage is dark. The living Swahili Muslim communities of Kilwa, Zanzibar, Pemba

and Mombasa are rendered with weight and as they understand themselves; Islam, the mosques, the Sufi

orders, the healing craft, and the Sayyid lineage as such are never the antagonist, and the

malevolent register is the coast's OWN condemned material — the ill-intentioned sea-spirit and the

harming branch of the craft the tradition itself sets against its healing branch. The Mijikenda kaya

forests are protected living sacred sites, referenced and never reproduced, and this shelf attests

NOTHING about them (Section 16 records that as an absence, not a silence). Substrate citations are

region-filtered (see Section 16). page_status: Research Phase.

Region Identity

• region_id: swahili_coast · Region name: Swahili Coast
• chapter_set: CH_08 (the join root) · Sub-regions: Kilwa Kisiwani and the Kilwa bay archipelago (Songo Mnara, Sanje ya Kati, Husuni promontory), the Mombasa and Zanzibar harbour waypoints, Pemba, and the Mijikenda kaya hinterland of the Kenyan south coast
• Modern country: Tanzania, Kenya, Zanzibar · Region center: 8.9572°S 39.5111°E (Kilwa Kisiwani); 6.1659°S 39.2026°E (Zanzibar)
• Time period: the Swahili trading-civilization peak approximately 1200-1400 CE, present-framed on the Mahdali century — the Husuni Kubwa palace of Hasan ibn Sulaiman (reigned 1320-1333) and the enlarged Friday mosque — with the earlier timber-and-daub town, the Portuguese and Omani centuries, and the conservation present reached forward and back across time-fluctuation windows · Era name: the ocean arc, the first crossing since the cave
• Season on arrival: the trade season — the monsoon regime the whole coast is timed to, arriving on one wind and leaving on the other (the region row's weather_cycle_profile cell is empty; Section 17 records that as a registry gap, not a page invention)
• Page status: Research Phase
• Last updated: 2026-07-27 · Sources processed this session: SRC_00006, SRC_00166, SRC_00172, SRC_00181, SRC_00184, SRC_00196, SRC_00198, SRC_00313

Character Quick Reference

• Protagonist entry point: landfall at the coral-stone island city after the first open-water crossing since Flores, apprenticed on the crossing to Nakhoda Binti Hassan; exit point: departure north by sea on the Sealed Letter's Aksumite pointer toward Ethiopia (Ch 9)
• Healer's child presence: carried, the silent-observation mystery still active and unexplained; Ntala (CHAR_0002) does not active-join until Ch 13 and takes no on-screen Swahili-coast beat here
• Velheim form for this region: the first fully-aware vessel, held entirely at design grain (NW_0007, Cat-5) — the player perceives a hostile recognition on sight and an unease they cannot name, never a family and never a pattern; see Section 11
• Key ally NPC: Nakhoda Binti Hassan (CHAR_0056), the dhow captain who teaches the Trade-10 navigation craft and never asks about the cloth wound over the ears — the first of the chapter's two positive recognitions
• Key enemy NPC: Sayyid Abdallah bin Rashid (CHAR_0055, BE_0008), the merchant who took the emptying bargain with open eyes — his corruption is his own and is never a failing of the honorific, the faith, or the community; the honoured Kilwa Sultan al-Hasan ibn Sulaiman (CHAR_0057) is a positive bearer of the same elite
• Neutral significant NPC: Li Wen (CHAR_0062), the porcelain merchant, and Mutota (CHAR_0060), the interior gold-route trader — the two ends of the ocean web met in one market
• Familiar available: Kasa the green sea turtle (W5) bonds on the reef run and Msafiri the coastal seabird (F3) bonds over the open water; Shell (W1) is sustained and Kira (F1), Khan (G1) and Muthu (G2) carry
• Faction alignment: no standing faction; the durable social writes are the Kilwa Sufi community's standing (FORK_CH08_A) and the coast's memory of the leaving (FORK_CH08_B), both read forward at the Ch-59 Africa continental-revisit

HEADER END — BODY BELOW

Zone Catalog

One row per zone. Rows sourced from spine §3b site_ids copy terrain / functions / ambient_threat /

vril_density / climactic_affordance / protected VERBATIM from the CH_08 site rows; the page-authored

columns are the page-side geometry (area_bounds, zone_bounds_wgs84) and build_priority plus the T0.2

extension-column pack. area_bounds are first-pass design sizings (meters, local ENU origin at the

region center), refined at the Section 1 terrain-build pass; they are page authority, not spine

canon. worldcover_biome, asset_routing and environment_grammar_refs are deliberately blank pending

their own passes, exactly as on the pilot page.

Zone-bounds derivation note. Four of the five envelopes are anchored on the region row's own Kilwa

centroid and on named ground the substrate places: the island town and its promontory palace sit

east of the modern Kilwa Kisiwani village, the palace itself on a barred spur at the top of a cliff

on the eastern promontory with a panorama over the whole bay [SRC: SRC_00184 — Kilwa dans Annales

islamologiques, p. 14]; the reef box takes the harbour approach and the reef crest, where the

substrate places the two recorded hazard concentrations — Matuso at the northern entrance to Kilwa

Kisiwani harbour, and the intertidal Jiwe la Jahazi wreck site on the sharp coral-limestone crest of

the east coast [SRC: SRC_00198 — Shipwreck Evidence from Kilwa, p. 13; p. 9]; the open-sea arena is

the water between the island and Songo Mnara, whose own coast the survey walked to the cape of

Kivurugo [SRC: SRC_00184, p. 9]. The three Kilwa-bay boxes are laid EDGE-ADJACENT and not overlapping —

the island's east parallel is the reef's west parallel and the reef's east parallel is the open-water

arena's west parallel — so intersection between every pair of boxes on this page is zero by area, and a

marker, encounter or gather node resolves inside exactly one zone. The kaya envelope is the one

page-authored placement with no

substrate under it at all: this shelf carries zero Mijikenda material (Section 16 records the

positive-controlled absence), so the box is a provisional coastal-hinterland footprint on the Kenyan

south coast, it identifies no actual grove, and it is explicitly NOT a reproduction of any real kaya.

It is re-derived at the build pass from the MUST-ACQUIRE Mijikenda source, and until then the build

treats the grove as an authored forest margin that borrows nothing from a specific living sacred site.

Care-tier column note. The care_tier column encodes PER-ZONE handling — the community-proximity of the

ground plus the care register of what is staged on it — and not the region tier, which is standard per the

spine's own CH_08 line and per Section 4. The two answer different questions and a consumer must not read

one off the other: the combat lane reads this column for the zone it is building, and Section 4 for the

region. Four of five zones read elevated for named reasons. kilwa_kisiwani is a living Muslim urban

community on a protected World Heritage island. mijikenda_kaya_edge is the protected sacred ground of a

living people this shelf attests nothing about, which raises rather than lowers the handling. pemba is a

real named living island community and the registry binds a harming-craft term to it, so the handling is

elevated even though the ground itself is unprotected. dhow_open_sea is open water and unprotected, and it

is elevated because of what is staged on it rather than what it is — the elevated-tier BE_0008 encounter

and the FORK_CH08_A hold, which is the chapter's slave-trade material at its closest range and is itself

declared elevated in Section 13. The encounter carries the tier there, not the terrain. Only

coral_reef_hazard_zones reads standard, because it is neither community ground nor an elevated-tier

encounter.

Zone-grain node-class note. The region row's node_class cell is EMPTY, along with the whole

environment-profile column pack (env_density_band_default, ambient_elemental_affinity,

weather_cycle_profile, terrain_feature_class_ids, movement_modifier, terrain_permanence_default,

terrain_integrity_skin, protected_terrain_ref, env_profile_status), where the flores_island and bali

rows carry values. That is a registry-population gap, not a page decision, and the page declines to

assert a node_class from inference. What the page can state on its own evidence is the shape the

values will have to respect: three of the five zones are protected or care-elevated and the dominant

register is a living Muslim urban community plus a protected sacred grove, so a transferable

successor-apparatus lane is not obviously present here and none is asserted. The registry pass owns

the cells.

Derived Indexes

Generated from the spine at build time; regenerate, never hand-edit. Every id below is a live join

into the typed layer.

• Boss index (every BE_xxxx staged in the chapter_set):
• BE_0008 "Sayyid Abdallah bin Rashid, the first fully-aware vessel" (CH_08, Thread 3) · verbs BOARD-AND-STRIKE/EXPOSE · axis 50/25/25 · phases P1→CH08_B09, P2→CH08_B10, P3→CH08_B11 · care tier elevated · dual win route, part-break to an HP finish or expose-pacify with the hold as evidence
• BE_0008_S1 the Reef-Shetani and the Pemba mchawi (BANISH/TRAVERSE, 30/45/25) → CH08_B06 — puzzle-forward; the win recovers the endangered crews and frees the reef
• BE_0008_S2 the Kaya-Forest Intruder (STRIKE/PROTECT, 40/25/35) → CH08_B07 — combat routes through the intruder only, never the Mijikenda and never the grove
• SQ index (every §11 row landing here):
• SQ6 pirate_naval — ADVANCE at Ch 8, the chapter's lead questline (beats CH08_B01, CH08_B05, CH08_B09); reward the naval and navigation depth, the Trade-10 kit, and the opt-in ownable mtepe
• SQ12 secret_societies — ADVANCE at Ch 8 (beats CH08_B08, CH08_B12); reward the antagonist-mystery intel with the pattern held from the player, and the Sealed-Letter mystery
• SQ1 fairy_tale_folklore — ADVANCE at Ch 8 (beats CH08_B04, CH08_B05, CH08_B06); reward the coast's own spirit-cosmology substrate and the harming-and-healing craft distinction
• Scene index (every §13 row bound to this region's beats):
• CH_08: SC_CH08_the_crossing, SC_CH08_reef_spirit_warded, SC_CH08_the_enemy_who_knows, SC_CH08_the_hold, SC_CH08_the_letter
• page-authored ambient-cinematic additions: none yet (the only scene rows a page may author; deferred to the build pass)
• Mark/state surface:
• marks written: the leaving mark (FORK_CH08_B) — the coast's memory of the protagonist, permanent on the generative and corrosive options, forward_track [CH_59]; the hold (FORK_CH08_A) writes no mark and rides the relationship and faction columns instead
• ws.* keys touched: ws.integrity_score, ws.identity_hiding_status, ws.reputation_biography_propagation, ws.relationship_history, ws.faction_standing_per_faction, ws.familiar_bond_state, ws.antagonist_operational_state_per_region
• Recall binding: recall node swahili_coast

The Twenty Numbered Sections

Section 1 — Geography and Topography

The Swahili coast is a low sedimentary shore of coral limestone, mangrove creek and fringing reef,

strung with islands that are harbours before they are anything else. The chapter's ground is the

Kilwa bay archipelago at the southern end of what the substrate calls the Zanj coast — a cultural

littoral running from Mogadishu in Somalia to the Bay of Sofala, between the mouths of the Zambezi

and the Limpopo [SRC: SRC_00184 — Kilwa dans Annales islamologiques, p. 2 n2]. Kilwa Kisiwani is an

island town; Songo Mnara, Sanje ya Kati, Songo Songo, Sikitiko and the Pandé peninsula ring the same

bay, and the Kiswere creek opens to the south [SRC: SRC_00184, p. 5; p. 10; p. 11; p. 29].

Four terrain facts from the survey record govern how this region reads underfoot, and each is a

gameplay affordance before it is scenery:

• The island is tidal, not sealed. At spring tides a ford at Serenge lets a person cross from the

mainland onto the western side of the island, coming out near the Makutani enclosure [SRC:

SRC_00184, p. 10]. A crossing that exists only at certain hours is a timing puzzle the coast hands

the designer for free, and it is the same timing discipline the monsoon teaches at ocean scale.

• The reef is the wall and the door at once. The way into the harbours of Kilwa Kisiwani and

Sangarungu is marked by reef-coral causeways, which themselves support the mangrove growth that

marks the reef edge; the archaeologists' reading is that the inevitable wrecks are what precipitated

the need to build them [SRC: SRC_00198 — Shipwreck Evidence from Kilwa, p. 13]. The safe lane is

literally a built thing, and it can be read, lost, or broken.

• The coral limestone bites. Hulls of the period were designed to flex over shallow reefs and sand

bars, but the sharpness of the coral limestone at Jiwe la Jahazi would have caused catastrophic

damage to a vessel carried across the crest by waves — and the island's own oral tradition tells of

a wreck there [SRC: SRC_00198, p. 9]. The reef zone's ambient_threat high is not a difficulty dial;

it is the ground.

• Landing places are specific and few. Ports on this coast prefer sand spits and land near inlet

mouths, for easy access and relatively deep channels; the alternative is a passage through the

mangroves reachable only at high tide [SRC: SRC_00196 — Settlement and Trade from AD 500 to 1800,

p. 17; p. 25]. Where a ship can touch is a hard constraint, and it is why a captain's knowledge is

worth apprenticing to.

Terrain data plus the build chain (researched terrain pipeline): DEM source Copernicus GLO-30 via the

OpenTopography API over the five zone bounding boxes; meters-per-vertex a per-region compression

decision (first-pass 1 m/vertex at the island town and 2 m at the reef and grove, 5 m over the open

water and the Pemba belt); the GDAL/rasterio recipe carves the harbour channel, the reef crest and

the Husuni cliff as scripted composites; a monsoon-shore erosion pass adds spit, creek and

tidal-flat detail; UE Landscape import tiles under World Partition at the 8,129-vertex actor cap; ESA

WorldCover maps mangrove, cropland and reef-flat landcover to material-layer weightmaps. No terrain

smoke has been run for this region yet — the pilot page records one for Liang Bua and this page

records its absence rather than borrowing those numbers.

Time-period modifications from modern: the present frame is the fourteenth-century town at its

height, so the Gereza fort (1807) and the Portuguese castle beneath it (1505-1512) are absent from

the skyline and are reached only through the Section-18 windows [SRC: SRC_00184, p. 17]; the mangrove

stands where the medieval town cut it, and the survey's own reading at Songo Mnara is that a

monumental quay built at the water's edge argues the mangrove was NOT there in the fifteenth century

[SRC: SRC_00184, p. 19; p. 27 n96], which the build honours by shifting the mangrove line between

strata rather than treating it as fixed. Climate and atmosphere: tropical-monsoon, humid, the

alternating trade winds that time the whole ocean, hard sun on lime-plastered stone, and the

seasonal reversal that decides when a ship may leave at all.

Vril sites: none rowed at Ch 8 in T0_Vril_Site_Registry. The coral-reef amplification substrate is

Thread-1 material catalogued forward rather than a recharge node, and baseline recharge carries from

the Ch-3 floor. The page does not mint a site the registry does not carry.

Section 2 — Flora and Botany

The region's botany is a working botany: almost everything the shelf attests about plants here is

about what people cut, plant, eat, burn or build with, and the page carries it that way rather than

inventing a wild flora it has no source for.

Mangrove is the region's structural plant. The substrate defines the ecosystem plainly — shrubs on

aerial roots and the larger trees behind them — and records that its intensive exploitation in the

medieval period is what makes a vanished mangrove line at Songo Mnara probable [SRC: SRC_00184, p. 27

n96]. Mangrove poles are gathered at landing places for export, a trade the excavated middens argue

is ancient, and the mangrove whelk in those middens shows the forest itself was worked for food as

well as timber [SRC: SRC_00196, p. 19; p. 26]. In game terms the mangrove is a gather node, a

building material, a boat-repair supply, and a maze — the creek passages that reach the sheltered

landing places run through it.

The coconut palm is the region's second working plant and it appears in the record at four different

jobs: as plantation ground on the islands, as leaves lining wells dug into sand where there is no

stone to line them with, as the fibre that mats and rope are made from, and as the fibre that stitches

a ship's planks together [SRC: SRC_00184, p. 27; SRC: SRC_00196, p. 25; p. 7; SRC: SRC_00198, p. 2].

That last is the one that matters most to Ch 8: the substrate's attestation of stitched planking is

Ibn Jubayr's record of vessels at the Red Sea port of Aydhab, carvel-built with planks stitched with

coconut fibre through holes bored at intervals near the edges of adjacent planks [SRC: SRC_00198,

p. 2]. The Trade-10 craft therefore has a real material chain behind it — palm to fibre to cord to

hull — and the cord is a gatherable, workable, spendable thing.

Rice is the region's grain and it is legible in the ground. The abundance of rice increases in the

later period at Kilwa, rice appears in ninth- and tenth-century deposits on the Comoros, and a whole

port on this coast is named for it — Muchelele, from mchele, uncooked rice in Kiswahili [SRC:

SRC_00196, p. 26; p. 24]. Millet is attested in its ritual frame rather than its agronomy: at

Makunduchi on Zanzibar the spirits and their religious experts had to be given their dues three

times within the context of millet cultivation, at the clearing of the field, when the cultivators

moved out to guard the growing crop, and before the harvest [SRC: SRC_00313 — the Makunduchi New

Year study, p. 3 n5]. Millet is thus the crop whose calendar is a ritual calendar, which is exactly

the shape a farming minigame wants.

A stratum warning the build must not lose. The one full plant list in the shelf is a modern survey

inventory from the Angoche islands in northern Mozambique — sweet potato, cassava, banana, maize,

coconut palm, papaya, sugar cane, pumpkin, mango and rice, with casuarina, cashew, baobab, tamarind

and citrus trees [SRC: SRC_00196, p. 25]. It is labelled as modern crops in its own source, it is a

different stretch of coast, and several of its entries — cassava, maize, papaya, cashew — are

post-Columbian and cannot stand in a fourteenth-century frame. The page therefore takes from it only

the pre-Columbian members that the record independently roots on this coast (coconut, rice, banana,

sugar cane, tamarind, baobab, citrus), carries them as plausible rather than attested-here, and

routes the fourteenth-century Kilwa archaeobotany to Section 16 as an acquisition target. The 2016

excavation floated every context for carbonised plant remains, so that archaeobotany exists and is

simply not in this shelf [SRC: SRC_00166 — Dating Kilwa Kisiwani, p. 9].

Build notes: foliage source-assets favour mangrove, coconut palm, casuarina and coastal thicket

matches; density high in the creek and the plantation belt, thinning to bare reef flat and

lime-burnt ground near the town; palette salt-green over grey coral rag, with the plaster white of

the town reading hot against it; wind steady and directional, because on this coast the wind is a

clock.

Section 3 — Fauna and Animal Life

The page CONSUMES the T0_Creature_Roster rows the spine already staged for CH_08 — the Swahili reef

cohort (CR_0043), coelacanth (CR_0044), dugong (CR_0045), whale shark (CR_0046), African fish eagle

(CR_0047), coastal seabird cohort (CR_0048), coastal dik-dik (CR_0049), yellow baboon (CR_0050), and

the Nile crocodile (CR_0060) — and FEEDS region-ambient detail. It mints no new rows.

The honest state of the evidence, stated first so nothing below reads stronger than it is: none of

the eight processed sources is a zoology or ecology work. There is no region-scoped natural-history

attestation on this shelf for the dugong, the whale shark, the coelacanth, the dik-dik, the baboon or

the crocodile, and the page therefore carries their behaviour as authorial general knowledge, bound

by the same real-ecology discipline the pilot page applies to its own fauna, rather than dressing it

in a citation it does not have. What this shelf DOES attest is the fauna the archaeology recovers,

which is a different and in some ways better thing: it is the animal life people actually met, ate,

avoided and depended on.

• Fish are the substrate of the diet and then stop being it. The Kilwa excavation reads a shift in

what the authors take to be the late twelfth century, from a diet rich in fish to one rich in

non-fish protein sources [SRC: SRC_00166, p. 14]. A town that gets rich changes what it eats. The

playable read is a market that sells differently in different strata, and a fishing trade whose

social standing moves under the player's feet across the Section-18 windows.

• The middens name the shellfish. Mangrove whelk (Terebralia palustris) and crown conch (Volema

pyrum), clam, cockle and olive shell (Ancilla), plus Nerites that normally live on intertidal rock

and were probably carried in with limestone or reef coral for burning [SRC: SRC_00196, p. 19]. That

is a complete, real gather table: mangrove forest, intertidal sand, seagrass, and rock — four

gathering grounds, each with its own species, one of them a by-product of the lime industry.

• Fish traps are built at the landing place. Modern fish-trap fences are still made where the boats

come in, and the shellfish are processed perhaps for bait and also for food [SRC: SRC_00196,

p. 13]. The trap is a placed, tended, returned-to structure — the shape a Trade-10 gather loop

wants rather than a swing-at-a-node loop.

• Fishbone is the loudest thing in the ground. A fourteenth- to sixteenth-century deposit yields a

large amount of fishbone alongside the mangrove-whelk shell, which the excavators read as showing

how important fishing was to the port [SRC: SRC_00196, p. 19]. Ambient design follows the evidence:

drying racks, gulls, bone middens, smell.

• Livestock and hives belong to the island economy. Sheep, goats, cattle and poultry, and bee hives,

are recorded on the island ground [SRC: SRC_00196, p. 25], and Songo Mnara still carries livestock

and plantations, notably coconut [SRC: SRC_00184, p. 27]. The satellite island as a supply farm for

the capital is the region's quiet economic fact, and it gives the player somewhere to be useful

that is not a market stall.

One animal in the shelf carries the whole ocean on its back. The record notes that Chinese porcelain

has been found as far inland as Zimbabwe, and that an African giraffe was once brought to China as a

gift to the Emperor [SRC: SRC_00006 — Kilwa Inventory Catalogue, p. 8]. That single sentence is the

region's best statement of its own reach, and the page carries it as a lore beat rather than an

encounter: a creature from the interior travelling further east than any person the player has met.

It also fixes the direction of the exchange honestly — this coast SENT, it did not only receive.

Familiar candidates — full entries, both consumed from the spine and neither re-authored here: Kasa

the green sea turtle (Water, slot W5, oceanic endurance), bonded on the open water during the reef

run, present whenever the Water slot is active thereafter and rotating forward across the Atlantic

and Caribbean crossings; and Msafiri the coastal seabird (Flying, slot F3, aerial reconnaissance),

bonded on the crossing itself, rotating forward through the African inland chapters and the Ch-59

revisit. Shell (W1) is sustained across both.

Folkloric beings route to Creature_Roster.folklore_attribution and none mints a BE_ row, since the

boss slate is spine authority. The one folkloric being the shelf actually attests is treated in

Section 5, under its own tradition's name and category, and the page deliberately does not stage a

second one from inference.

Section 4 — People and Culture

The single most load-bearing sentence this shelf gives the page is a definitional one, and it governs

every other line of this section: the Swahili are not one people but a myriad of ethnic groups united

by a common culture based on Islam, an African culture of the East African shore running from

Mogadishu in Somalia to the Bay of Sofala [SRC: SRC_00184, p. 2 n2]. Everything the build does here

is bound by that. There is no single Swahili costume, no single Swahili face, no single Swahili

speech, and a crowd in this market is a crowd of many origins holding one culture in common. §17.1

binds throughout: these are living communities, rendered with weight and as they understand

themselves.

The second load-bearing fact corrects a specific and still-circulating error. The idea that Swahili

towns were founded by immigrants, or that immigrants made up the majority of their populations, has

been widely demonstrated as false through a series of studies of urban origins and evolution along

the coast [SRC: SRC_00166, p. 12]. The Kilwa Chronicle's own Shirazi foundation story is carried in

this world as what it is — a real, living, politically live origin tradition that the community tells

about itself, with the archaeology of an African town that grew where it stood underneath it. Both

are true things of different kinds, and the page holds them together rather than making one debunk

the other. The record itself supplies the reason the chronicle is not naive: oral traditions along

this coast functioned to strengthen community identity and to legitimise the local order at a time

when the coast was ruled by European and Omani powers [SRC: SRC_00196, p. 5]. A founding story is a

political instrument as well as a memory, and a player who learns that has learned something real.

Social structure, as far as the shelf carries it. The word for the notables is attested: wa-ungwana,

in Kiswahili, designates the notables or patricians who occupy the large stone houses in the centre

of the towns [SRC: SRC_00184, p. 22 n93]. The same record immediately undercuts the easy reading — a

large stone house testifies to the presence of a wa-ungwana notable, but the building says nothing

about the importance of the site it stands in, and Swahili stone settlements are not always important

centres [SRC: SRC_00184, p. 27]. Stone is a statement about a household, not a ranking of towns. On

Zanzibar the shelf attests a descent-based structure of a different kind: at Makunduchi descent is

reckoned both matrilineally and patrilineally, so a person can belong to several descent groups at

once, and each descent group carries its own inherited relations with local spirits — which is

described as a strong marker of local identity [SRC: SRC_00313, p. 2]. Traditional political

leadership there was closely linked with religious expertise and consisted of a council of five

elders [SRC: SRC_00313, p. 3].

Language and literacy are the region's own achievement and the record states the mechanism.

Kiswahili grew out of contact between the Arabic-, Persian-, Indic- and Indonesian-speaking peoples

of the northern and eastern Indian Ocean and the coastal peoples of East Africa; the northerners

introduced writing, and formal education in the madrasa spread literacy to the masses with the

arrival of Islam [SRC: SRC_00172 — Early Globalization in the Indian Ocean, p. 8]. The Arabic

loan-layer is specific rather than diffuse: south Arabian contact shows particularly in the

vocabulary of the parts of a ship and of shipbuilding, most of all among the northern Swahili

geographically closest to the Hadramaut coast and most mixed with Hadrami people [SRC: SRC_00172,

p. 8]. That is a beautiful, usable fact — the trade's technical vocabulary is where two languages

actually met, and the Trade-10 lessons are exactly where the player hears it.

Appearance and dress. The shelf's one description of how people on this coast dressed is Portuguese

and sixteenth-century and about Angoche rather than Kilwa, and the page marks it as such rather than

promoting it: Barbosa described the people of Angoche, some of whom were Arabic-speaking, as having

black or tawny skin, wearing silk and cotton cloths below the waist with other cloths folded like

cloaks over the shoulders, some wearing turbans and others caps made of squares of silk cloth [SRC:

SRC_00196, p. 7]. Two independent lines corroborate the centrality of cloth without dating it: the

Kilwa Chronicle's first sultan is remembered as having obtained the island in exchange for lengths of

cloth and is nicknamed nguo nyingi, the man of many clothes [SRC: SRC_00184, p. 3 n7-n8]; and the

pillar of a Swahili tomb is shaped so that its upper swelling represents a stylised turban, the

headgear that was formerly hung at the top of Islamic graves [SRC: SRC_00184, p. 15]. Cloth is

currency, cloth is rank, cloth is what a grave says about a man. The fourteenth-century Kilwa

costume itself is an honest gap and Section 16 carries it.

Builders of knowledge — the L2 civilizational-marks lane, each credited to the record that attests it:

• Swahili maritime urbanism in coral. A whole architecture of coral rag, lime and plaster, with

domes, barrel vaults, fluted semi-domes and arches carried without keystones by the weight of their

own lateral counter-thrusts [SRC: SRC_00184, p. 16; p. 13]. That is structural engineering solved

in a material almost nobody else built monumentally with, and Section 7 carries the detail.

• The Husuni Kubwa palace as considered design. The record's own judgement is unambiguous: a

monument of a grandeur and quality never equalled on the African coast, and a true Swahili palace

with a thought-through plan rather than an accretion of large houses around a central court, as at

Songo Mnara or Gedi [SRC: SRC_00184, p. 14]. Architectural intent, credited as intent.

• Coinage. The excavators found a local production of silver and copper coinage repeating some of the

names in the Kilwa Chronicle [SRC: SRC_00166, p. 12], and the sultanate had begun minting its own

currency in the middle of the eleventh century [SRC: SRC_00172, p. 1]. A state that mints is a state

that guarantees, and this one guaranteed far enough that five of its copper coins turned up in the

Wessel Islands off northern Australia [SRC: SRC_00172, p. 1].

• Monsoon navigation as applied astronomy and arithmetic. Ahmad ibn Majid's fifteenth-century Poem of

Sofala carries recognisable place names and approximate sailing distances along the East African

coast, giving the length of Zanzibar as one day's sailing and implying a day of roughly eighty to a

hundred kilometres — with the record immediately noting that real daily distance varied with the

direction and strength of the monsoon and the currents, the possibility of night sailing, the need

to avoid wreck hazards, and the demands of trading and servicing the vessel [SRC: SRC_00196, p. 4].

That is a working pilot's arithmetic with its own error terms stated, and it is the Trade-10

navigation mechanic's grounding.

• The sewn ship. The mtepe, an east African sewn-plank ocean-going cargo vessel, is attested here in

its own people's images rather than in a European description — a fifteenth-century engraving in the

Barrel-Vaulted House at Songo Mnara shows its swan-necked bow with a pennant flying from the mast

[SRC: SRC_00198, p. 3]. A shipbuilding tradition that leaves its portrait on its own walls.

• Literacy and the manuscript world. The madrasa network and the Arabic script that carried it [SRC:

SRC_00172, p. 8; p. 11], and the coast's own written history in the Kilwa Chronicle, whose dynastic

sequence the archaeology can be laid against name by name [SRC: SRC_00166, p. 12].

• Heritage conservation as a living craft. The 2002-2004 Kilwa works were run as a training site

whose explicit purpose was to train qualified Tanzanian workers and transmit knowledge of the

techniques for conserving coral-limestone ruins [SRC: SRC_00184, p. 11]. That is a modern

achievement of the same community and it is credited in Section 19, not folded into the medieval

frame.

Festival and daily life — the immersion lane, every beat opt-in and non-gating. One festival is

attested on this shelf, in detail, and it belongs to a specific place and a specific stratum, both of

which the page keeps: the New Year at Makunduchi in southern Unguja, documented in fieldwork of the

early 2000s [SRC: SRC_00313, p. 3; p. 11]. Its shape as recorded: two weeks before the day, the Koran

is read at some of the mizimu that guard the borders of the town, and then the town is circled while

Allah is praised, both understood to drive away evil spirits, and an animal is sacrificed; a week

before, the spirits of the important mizimu receive sacrifices of food and cloth, because that is when

they are believed to celebrate their own New Year; on the day itself a hut is built in a central area

and set alight, and as it burns it is circled by women singing, and the smoke is read to predict the

coming year's weather; the young men of the southern and northern halves of the town fight each other

armed with sticks; and in the evening there are dances [SRC: SRC_00313, p. 3]. The page carries this

as the region's calendar-festival anchor with three binding conditions. It is Makunduchi's, not

Kilwa's, and is never relabelled as a general Swahili rite. It sits in the Section-18 modern window

rather than the fourteenth-century present, since the fieldwork is contemporary and the source itself

traces the festival's twentieth-century political reshaping. And the ceremonial content is referenced

rather than reproduced: the player may attend, may take a role in the stick contest, may carry cloth

to a sacrifice, and the rite's own words are not authored for it.

The everyday layer the shelf supports is the port's. A port of call exists to give ships water, food

and firewood and to repair them, and some ports of call were used for religious reasons, for

offerings or prayer at a mosque or a cave for a safe voyage [SRC: SRC_00196, p. 25]. That is the

texture of the market beat: provisioning, repairing, and praying before a crossing, in that order,

because they are the three things a departure needs.

Represent-with-care posture. The Swahili Muslim communities of Kilwa, Zanzibar, Pemba and Mombasa

and the Mijikenda are protected communities; combat never routes through them; Islam, the Great

Mosque, the Sufi orders, the healing craft and the Sayyid lineage as such are never antagonists, and

the honoured Kilwa Sultan is a positive bearer of the same sharifian elite Abdallah dishonours. Care

is calibrated rather than inflated: the region renders at care tier standard, the tier the spine

stages the chapter at, with two strands carrying elevated handling wherever they are reached — the

Indian-Ocean slave trade, and the harming-and-healing craft material, because both touch living

communities and living reputations. Care means weight and authenticity, never omission. The hard

material is carried: the fourteenth-century town levied taxes on a trade that moved people as well as

gold, an eighteenth-century residential enclosure on this island played an essential role in the slave

trade and a mainland road ran to it through the Serenge ford [SRC: SRC_00184, p. 10], and the record

plainly says the slave trade and successive occupations left scars and destruction on buildings,

people and the landscape [SRC: SRC_00006, p. 8]. What §17.1 protects is the collective: no living

people cast as villain, defamed, or made the target. It does not make an individual antagonist who

belongs to that community untouchable, which is exactly why the boss slate routes where it does.

Section 5 — Spiritual Traditions and Occult

Cosmology. This is a Muslim coast and has been for a very long time — the earliest concrete evidence

of Islam in eastern Africa is a mosque foundation at Lamu with gold, silver and copper coins dated

830, and the oldest intact building in eastern Africa is a functioning mosque at Kizimkazi on southern

Unguja dated 1007 [SRC: SRC_00172, p. 3]. Islam arrived through traders and settlement rather than

conquest and remained an urban and coastal phenomenon for a long time [SRC: SRC_00172, p. 4]. By the

time of the traveller's visit in the fourteenth century almost every coastal settlement from Somalia

south was Muslim, and Arabic was the common literary and commercial language of the whole ocean [SRC:

SRC_00172, p. 4]. The faith is the water this region swims in, and it is never staged as exotic,

never as veneer, and never as the thing the protagonist has to see through.

Inside that, the shelf attests one detailed local spirit-cosmology, and the page states its scope

precisely before using it. It is Makunduchi's, in southern Unguja on Zanzibar, recorded in

contemporary fieldwork. In it, spirits commonly called shetani, plural mashetani, reside in specific

places called mizimu, singular mzimu, which are marked by natural features — caves, stones, trees or

thickets — and small huts are often built for them; ritual interaction happens at the mizimu, and the

main feature of the practice is the sacrifice of food or cloth [SRC: SRC_00313, p. 1]. The relation is

reciprocal: the spirits are given their due, and in return they secure the fertility of the land and

the people and aid their human relations in sickness and personal trouble [SRC: SRC_00313, p. 2].

Every citizen inherits relations with local spirits through their descent group, so belonging to the

town and belonging to its spirits are the same fact [SRC: SRC_00313, p. 2].

The tradition's own moral architecture is the part Ch 8 is built on, and it is attested exactly. From

local perspectives these practices are classified as mila — local, traditional, rural custom — as

opposed to dini, orthodox Islam; yet Islamic elements such as prayers to Allah and recitation of the

Koran are part of most of them, and in Makunduchi most people take part in the mila practices and

still consider themselves good Muslims [SRC: SRC_00313, p. 2]. Representatives of reformist Islam

criticise the spirit rituals as idolatrous, which the page carries as a live internal debate of the

community rather than as a verdict [SRC: SRC_00313, p. 2]. Within that field the rituals of the mizimu

count as uganga, religious practices of healing, as against uchawi, religious practices of harming —

and the record is careful about how clean that line is: it is generally suspected that a practitioner

of uganga can practise uchawi as well, there are no public accusations and nobody admits to uchawi,

and both employ the same means to opposite ends, namely plants, spirits, and the Koran, with the Koran

thought to be the most effective in both cases [SRC: SRC_00313, p. 2]. That is the region's whole

antagonist logic in one paragraph, given by the tradition itself: harm and healing are two branches of

one craft, not a foreign evil against a native good, and the same means serve both. The chapter's

sub-boss stands exactly there — the harming branch, condemned by the tradition that names it, faced

alongside the healing branch that opposes it.

The religious experts are named and their access rule is stated: within every descent group, experts

called wavyale, singular mvyale, take care of the rituals directed towards the spirits; most are

elders because expertise takes time to acquire, and both men and women can be religious experts [SRC:

SRC_00313, p. 3]. In the past the wavyale controlled access to land through the rituals required for

appropriating local resources, and the spirits and their experts had to be given their dues three

times in the millet cycle and again in connection with hunting and with net and trap fishing [SRC:

SRC_00313, p. 3; p. 3 n5]. Permission to use ground is a ritual transaction — which is a mechanic, not

a flourish, and it is the honest local form of the assimilation-integrity system for this region.

Two further attested facts keep the field from reading static or closed. It was possible to invent new

mizimu and to challenge a senior expert's position within a descent group; leadership was competed for

[SRC: SRC_00313, p. 6]. And a stranger could rise: the record carries the case of a man who came to

Makunduchi from the African mainland in the middle of the nineteenth century, received permission to

settle from the elders of the town, began in a subservient position, and went on to gain a leading

role without any connection to a local mzimu at all [SRC: SRC_00313, p. 6]. That is the page's

strongest single grounding for the arc's outsider mechanic — the door to standing in this world is

permission from the elders followed by service, and it is a documented door.

Ritual sites in the Kilwa district itself are attested at a coarser grain but attested. The record

states that the ritual dimension is an integral part of the district's culture, that these sites are

highly respected by the population, that people go to them to ask for healing, to obtain protection

and to seek support from the spirits in their activities, and that some are linked to legends — the

example given being the Bangwe stone, said to be a petrified imam who did not comply with Koranic

rules [SRC: SRC_00006, p. 77]. The page takes the Bangwe stone as a real, named, cited local legend

and stages it as a discoverable site with a story attached, exactly as the record frames it, with the

imam's story kept as the community's own cautionary tale and never played for mockery of the faith.

The dead are architecture here, and the architecture says who they were. Pillar tombs are very

frequent on the Swahili littoral; the pillar signifies the presence of an important person — imam,

emir, sharif or sheikh — and one recorded pillar is formed of two superposed shafts of decreasing

diameter capped by a swelling that represents a stylised turban, the headgear formerly hung at the

top of Islamic graves [SRC: SRC_00184, p. 15]. A whole small mosque could be raised as a memorial: the

best-preserved building at Kilwa Kisiwani, the domed mosque of the mid-fifteenth century, is read as a

funerary mosque built in memory of an important person, on the evidence of the pillar surmounting its

central vault, the pillar always being associated with funerary structures and deriving from a Seljuk

Persian or Fatimid Egyptian Islamic tradition [SRC: SRC_00184, p. 16]. At Songo Mnara the isolated

hill mosque holds at least sixty burials in its courtyard, its mihrab abnormally turned to the

north-west, and it stands visible to every boat navigating the bay [SRC: SRC_00184, p. 18]. Read a

skyline on this coast and you are reading a list of the dead.

The sea has its own devotional practice and it is the region's loveliest playable detail. Ship

engravings are found on the outer walls, outer windows and door jambs of mosques, and near the main

entrances of domestic buildings, and the archaeologists who first catalogued them believed them to be

charms or votive offerings for the safety of the vessel depicted [SRC: SRC_00198, p. 2]. Three

fifteenth-century engravings from the Great Mosque of Kilwa show curved flared bows carrying oculi and

tasselled and bobbled amulets, with yards supporting square sails and a spreader yard at the foot of

the sail [SRC: SRC_00198, p. 3]. A ship with eyes and charms hung on it, cut into the wall of a mosque

by someone who wanted it to come home. The player can find them, read them, and — at the Trade-10

tier where it is earned — cut one.

Naming discipline (the binding rule for this page, and the reason several familiar words do NOT appear

above). A literal term census was run across all 472 chunks of the region-filtered shelf, positive-

controlled against terms known to be present, which return in the dozens. The following return ZERO

occurrences: waganga, mganga, ngoma, jini, majini, Qadiriyya, Shadhiliyya, Ajami, taarab, tendi,

nakhoda, mchawi. Shetani returns once, in the Makunduchi passage cited above. What follows from that,

applied strictly:

• The spine's practitioner term, waganga, and its warding rite, the ngoma ya majini, are spine canon

and are consumed as such — this page does not contradict them and does not re-author them. But they

are NOT attested anywhere in this shelf, and the spine roots them in Mackenrodt and Giles, neither of

which is in the processed corpus. So the page carries the healing practitioners descriptively wherever

it writes region-scope substrate of its own, uses uganga and uchawi where the record actually attests

them, and routes the term-and-rite verification to the Section-16 acquisition. Supplying an unattested

local word inside living-tradition substrate is precisely the authenticity failure §17.1 guards against.

• The sea-spirit's local name is likewise not settled by this shelf. What the shelf attests is the

category — shetani, plural mashetani, the spirits of place — and it attests it for Makunduchi.

The chapter's reef-spirit is therefore carried in the tradition's own attested category with its

attested plural, described rather than given a species-name this record does not supply.

• The Sufi orders the spine names at Kilwa are canon and are consumed; this shelf attests the Sufi

network only in a modern bibliographic reference to Islamic Sufi networks in the western Indian Ocean

[SRC: SRC_00196, p. 28], which is a pointer and not an anchor. The order names go to the acquisition

list rather than into any player-facing string on this page's authority.

• The tendi flag. Thread 1 makes the Swahili tendi the third independent vital-force parallel after the

Batak and the Vedda, and that assignment is thread-doc and spine canon which this page consumes

without alteration. But the word occurs zero times in this shelf and no vital-force vocabulary is

attested here at all, so the page mints no gloss, no etymology, and no player-facing use of the term,

and the verify entry stands in Section 16 exactly as its sibling stands on the Sri Lanka leg.

• The two CONSUMED terms, nakhoda and mchawi, declared here so no reader mistakes them for shelf-attested

vocabulary. Both return zero occurrences at word boundary across the 472 chunks and zero again on a

naive substring pass, and the variants nahodha, nakhuda and wachawi return zero as well, against a

positive control on the same census returning wa-ungwana 3, kokoto 2, uganga 4, uchawi 5, wavyale 4,

mtepe 3, mila 7, dini 4, mchele 1 and Husuni 37. Neither term is supplied by this page. Nakhoda arrives

with the chapter's named ally, Nakhoda Binti Hassan (CHAR_0056), rowed in T0_Character_Index and carried

nine times in the spine entry; mchawi arrives inside the boss_name of BE_0008_S1 in

T0_Boss_Encounter_Registry. The page consumes both unaltered, re-authors neither, and routes both to the

Section-16 verify list against the same Mackenrodt and Giles acquisition that would settle waganga. The

mchawi case carries the sharper obligation, because the registry row binds the term to Pemba, a real

named living island community: no player-facing string on this page's authority labels Pemba a sorcery

island, this section handles the harming and healing branches as one coast-wide craft rather than a

place's property, and the Pemba attachment rides the registry row alone.

The vital-force parallel itself (Thread 1) is carried where it belongs: as substrate the player feels

through the recharge and the reef rather than through vocabulary, and the firewall against the

Bulwer-Lytton and "Vril Society" register is absolute — vril reads mathematical and physical, never

race-mystical, and nothing on this coast is staged as a lost superior science.

Section 6 — Weapons Military and Conflict

Era weapons and the martial layer. The chapter anchors no weapon: T0_Weapon_Registry carries no

canonical chapter acquisition at Ch 8, and the first Brute chapter-anchored add lands at Ch 9-10. The

martial surface here is therefore the naval one — boarding, deck fighting, rigging hazard, and the

recombination of every verb the arc has taught since Ch 2 — which is what makes Ch 8 the arc's first

Tier-A exam rather than a new-toy chapter.

Fortification is this coast's real military archaeology and the shelf is strong on it. Kilwa's first

sultan is recorded as fortifying his town not only against the mainland but against the other Swahili

settlements of the region [SRC: SRC_00184, p. 3]. From 957 to 1131 Kilwa fought its neighbour Shâgh

for control of the region, prevailing with difficulty in the twelfth century and taking the monopoly

of the Sofala gold trade [SRC: SRC_00184, p. 3]. The likely Shâgh is on the ground: Sanje ya Kati was

walled on a perimeter of 110 by 340 metres, which matches the chronicle's description of a powerfully

fortified town, and it was occupied from the tenth to the early thirteenth century — dying, in other

words, about when Kilwa won [SRC: SRC_00184, p. 29]. Sultan Suleiman b. al-Hassan (1170-1189) renovated

the city and built the fortified caravanserai of Husuni Ndogo, a quadrangular enclosure flanked by

towers, dated 1191-1215 [SRC: SRC_00184, p. 3; p. 14]. And at Songo Mnara the defence is the town

itself: the blind façades of its stone houses formed an effective rampart [SRC: SRC_00184, p. 18]. A

wall you live inside is a different design object from a wall you man, and the build treats it so.

Documented conflict, staged by stratum:

• The Kilwa-Shâgh wars (957-1131) and the fall of Sanje ya Kati. A war between two Swahili towns for

control of a trade, ending in one town's disappearance from the record. Reached through the

Section-18 deep window. Neither town is staged as the villain; what the surface asks is what a

monopoly costs, which is the chapter's own question in another key.

• The succession civil wars of 1466-1476, which shook Kilwa and ended with vassal cities such as

Sofala and Zanzibar emancipating themselves [SRC: SRC_00184, p. 3]. A hegemony coming apart from the

inside, reached forward through the windows.

• The Portuguese century. The Portuguese fort of Saint-Jean was erected at Kilwa in 1505 and dismantled

in 1512 [SRC: SRC_00184, p. 17]; the sultanate was nearly destroyed by the Portuguese in 1505 [SRC:

SRC_00172, p. 1]. The record also carries the other direction: a Portuguese ship grounded on a sandbank

off Kilwa in 1519 and its crew of ten, all but one youth, were killed by people from Kilwa, Mafia and

Zanzibar [SRC: SRC_00198, p. 26]. Both are in the record and both are carried; the coast is not staged

as a passive object of somebody else's century.

• The Omani stratum. The Gereza fort's foundation inscription, deciphered in 1857, gives 2 April 1807;

its plan is a quadrangular enclosure with two diagonally opposed circular towers characteristic of

Omani military building, a single gate in the middle of the east façade, a three-storey square tower

at the south-east angle, and a circular well 3.8 metres deep in the court — and the partial collapse

of its north-east tower revealed the Portuguese spur-bastion built into it, a fort inside a fort [SRC:

SRC_00184, p. 17]. A single ruin containing two empires is a set-piece the build should not waste.

Ritual combat, in its own register and not confused with war: at the Makunduchi New Year the young men

of the southern and northern halves of the town fight each other armed with sticks [SRC: SRC_00313,

p. 3], and in the post-revolutionary period fighting was permitted again with the weapons restricted to

the stems of banana leaves [SRC: SRC_00313, p. 11]. That is a bounded, scored, communal contest with

real rules and real stakes of pride, seated in the Section-18 modern window with the rest of the

festival, and it is the region's one attested martial-training surface. Its SQ2 potential is recorded

here and deliberately not typed to a stage, because the chapter's SQ slate is SQ6, SQ12 and SQ1 and the

page does not add a stage the spine does not carry.

Legendary candidate: none is proposed from this region. The 72-weapon roster is closed at 36 Brute and

36 Trickery with all four to-add slots resolved, this chapter anchors no acquisition, and the honest

naval material here is a technique set rather than an object. Proposing a legendary blade for a coast

whose martial record in this shelf is architecture and seamanship would be invention, and the page

declines it.

Acquisition standing: this pass does not close the martial gap. There is no region-scoped attestation

of Swahili personal arms, drill, or the boarding fight in the processed corpus, so the Section-16

verify-list entry stands, and the boarding set-piece is built from naval physics and the attested rig

rather than from a weapons record the page does not have.

Section 7 — Architecture and Engineering

REAL cultural architecture is reference-built, never raw-generated (§17.1); generation is sanctioned

for fantastical and vril structures only. Every monument named here is a protected UNESCO-listed

heritage site and is built as a type-accurate referenced composition, never a high-poly replica.

This is the section the shelf is strongest in, and the region's built environment is a documented

achievement of the first rank.

The material and the method. The buildings are coral: cut coral rubble laid in regular linear courses,

with sea-coral elements — quarter-colonettes, hemispherical bosses, finely cut miniature mihrabs —

used as dressed decorative stone [SRC: SRC_00184, p. 29; p. 16]. The aggregate under the floors is

kokoto in Kiswahili, coral-limestone pebbles under five centimetres across, obtained by hammering

madrepore blocks [SRC: SRC_00184, p. 21]. The binder is lime, and the lime is burnt on site — the

Kilwa sequence contains four superimposed phases of lime kilns calibrated between 1022 and 1329 [SRC:

SRC_00166, p. 13]. Reef coral, lime and limestone were traded along the coast to places that had none

[SRC: SRC_00196, p. 27]. That is a complete industrial chain the player can walk end to end: quarry the

reef, burn the kiln, hammer the kokoto, lay the course, plaster the floor.

The structural achievement. In the domed mosque the prayer hall is divided into three bays and three

naves by four octagonal piers on quadrangular bases; the pointed two-moulded arches attach to the walls

through pilasters crowned by rectangular capitals; the arch tops stand 2.5 metres above the floor; and

there is no keystone — the arches are held by the weight of their lateral counter-thrusts [SRC:

SRC_00184, p. 16]. The central dome rests on four rows of corbelled cornices and is ornamented with

three concentric circles of ceramic bowls set into its vault, flanked by two semicircular vaults also

carrying ceramics in the masonry, so as to open height over the mihrab [SRC: SRC_00184, p. 16]. In the

Great Mosque, two of the easternmost domes carry fluted vaulting while eight bays on the second and

fourth aisles are roofed with barrel vaults, and its earlier flat roof was carried on nine polygonal

single-piece columns whose bases are still visible in the floor [SRC: SRC_00184, p. 13; p. 12]. Thread

13 anchors here and it anchors on real geometry: bay, nave, dome, counter-thrust, and a decorative

programme in which imported porcelain is a structural ornament.

The palace. Husuni Kubwa sits on a promontory east of the island on a barred spur at the top of a

cliff, sited where it is windy and commands a panorama of the whole bay, and dated by a coral-stone

dedication to Hasan ibn Suleiman, who reigned from 1320 to 1333 [SRC: SRC_00184, p. 14]. It divides

into two functionally distinct complexes, one commercial and one princely. The southern part is a great

square court ringed with storerooms, so the sultan could warehouse the goods transiting Kilwa and levy

his tax on them; the main entrance is at the south-west toward the town, a wide stair rising to a

platform that overlooks the court. Its most characteristic elements are a large sunken audience court

ringed with steps and ornamental niches; the sultan's principal pavilion set to catch the prevailing

wind; sea-facing apartments richly decorated with mouldings and vaults of cut coral; two long rooms

under barrel vaults with panels and friezes; an octagonal pool framed by niches and benches west of the

living rooms; and, at the northern end, a monumental stair cut into the living rock descending to the

sea, with a small mosque at its foot flanked by two basins [SRC: SRC_00184, p. 14]. A palace whose

front door is a staircase out of the ocean is a level design, and the build should treat it as one.

Water engineering, at three scales. The mosque's ablution court holds two large rectangular basins

whose internal angles are reinforced with quarter-colonettes of sea coral, fed by a conduit that brings

water from a well outside the building [SRC: SRC_00184, p. 16]. Domestic sanitation is deeper than the

religious architecture is complex: Swahili house latrines can reach six or seven metres, and the

excavators' own observation is that domestic arrangements at Songo Mnara were more complex than the

construction of its religious buildings — with one latrine abandoned and its space reused before the

end of the fifteenth century [SRC: SRC_00184, p. 24]. And where there is no stone, wells are dug into

sand and lined with coconut leaves [SRC: SRC_00196, p. 25]. Three solutions to water at three budgets.

The harbour works. At Songo Mnara a massive double-faced wall with rubble fill was built right at the

waterline, sheathing the foot of the cliff and continuous with the enclosure to the north, with a

monumental yellow-sandstone staircase at the north giving onto it and a second stair leading up to the

funerary mosque — a landing zone for passengers and goods, and the reason the survey reads the western

part of the site as its disembarking quarter [SRC: SRC_00184, p. 19]. Out on the water, the reef-coral

causeways mark the reef edge and the way into the harbours [SRC: SRC_00198, p. 13]. Engineering that

tells a stranger where the sea is safe is the most humane kind, and it is the physical form of the

chapter's navigation puzzle.

Engineering feats and Book-of-Trades entries: the lime-burn cycle, the kokoto aggregate preparation,

the counter-thrust arch, the causeway build and the quay revetment are the Trade-8 Engineering

substrate seeds for this region (echo-seeds; the Trade-8 home roots elsewhere in the arc). The agility

course for the chapter is the mtepe dhow-rigging traverse, biome-rotated against the chamber-and-forest

courses of Ch 7 — the vertical is cordage and spar over open water rather than stone or canopy.

Section 8 — Politics Territories and Faction Relationships

Political structure: a sultanate that is a commercial instrument. The chronicle sequence the shelf

carries runs Ali b. al-Hussein as first sultan of Kilwa, with chronicle dates in the range 957-985;

Dawûd b. Sulaiman (1131-1170), under whom Zanzibar, Mafia and Pemba became vassals of Kilwa; Suleiman

b. al-Hassan (1170-1189), who renovated the city and built Husuni Ndogo; the Shirazi dynasty ruling

until 1277, replaced by the Mahdali dynasty out of the Hadramaut; and al-Hassan b. Sulaiman II, who

enlarged the Friday mosque and built the palace at Husuni Kubwa [SRC: SRC_00184, p. 2; p. 3]. The

archaeological frame agrees on the shape and revises the joints: the earliest coins and stone buildings

belong to a period the excavators date 1200-1290, and the Mahdali coins, the Husunis and the Great

Mosque extension to about 1290-1400, with the traveller's visit in 1331 inside that band [SRC:

SRC_00166, p. 13]. The present frame of this page sits in that last band.

Territory is a network, not a border. Sofala was a vassal of Kilwa and traded with the kingdom of Great

Zimbabwe, whose prosperity the gold profits materially served and which worked many mines on the high

plateaus of Sofala's hinterland [SRC: SRC_00184, p. 3 n13]; the Kilwa Chronicle records Kilwa gaining

control of Sofala in the thirteenth century [SRC: SRC_00196, p. 26]; Zanzibar, Mafia and Pemba were

vassals from the twelfth [SRC: SRC_00184, p. 3]. What Kilwa held was a chain of tolls and permissions

along a route, and the palace's storeroom court is where that abstraction becomes a room.

Power is also negotiated at street level, and the shelf gives the two mechanisms. Kilwa was a state

that could tax: the sultan warehoused transiting goods and took his share of them [SRC: SRC_00184,

p. 14]. And a port state of this coast was, in the analysis the record advances, primarily the fruit of

an alliance between foreign merchants and local authorities, which once established controlled both

the local distribution of luxury goods and the long-distance trade [SRC: SRC_00181 — Global Trade and

Swahili Cosmopolitan Material Culture, p. 45]. That is the exact political shape Ch 8's antagonist

exploits: a wealthy merchant with the sultanate's tolerance is not an aberration of the system, he is

the system's most powerful private actor. It is also why the fight can be won by exposure — in an

alliance polity, evidence is a weapon.

Five-Factors faction table (per neighbour). Path language uses the ratified fork vocabulary: options

are ACTION-keyed with direction {generative, neutral, corrosive}; bands are DERIVED, never option

labels.

• Sayyid Abdallah's house and its dependants (BE_0008, NW_0007): relationship type hostile (a private

commercial apparatus, not a community); player impact — the CH_08 main line resolves it by toppling

or exposing; standing on arrival adversarial from the moment of recognition; improve n/a; questline

the CH_08 main quest and SQ12

• The Kilwa Sufi community around Scholar Amara (CHAR_0059): relationship type protected-host and

refuge-keeper; player impact set by the hold (FORK_CH08_A) — freeing the captives raises standing

because the community is the refuge that receives them; standing on arrival hospitable-to-a-stranger;

questline SQ12 and the Sealed-Letter line

• The Kilwa sultanate and its court (CHAR_0057): relationship type protected-authority; player impact

light and indirect; standing on arrival formally neutral, privately informed; improve by conducting

oneself well in the market and at the mosque; degrade by giving the court a reason to notice a

stranger; questline none of its own at Ch 8

• Nakhoda Binti Hassan's ship and crew (CHAR_0056): relationship type ally-teacher and trade home;

player impact set by the leaving (FORK_CH08_B), which decides whether her Ch-9 and Ch-38 crossings

are available; standing on arrival earned by work; questline SQ6 and TRADE_10

• The waganga healers of the coast (un-named at spine scope): relationship type protected-allies; player

impact set by siding with them at the reef; standing on arrival wary and then warm; questline SQ1

• The Mijikenda of the kaya hinterland: relationship type protected-community; player impact set by

securing the violated edge; standing on arrival distant, closing to positive recognition; questline

the CH08_B07 line only — and the page asserts no further Mijikenda content, because it has none

• The ocean web (Mutota CHAR_0060, Li Wen CHAR_0062): relationship type neutral-economic; player impact

light; improve by trade and by carrying news honestly; questline the market and rumour chains

Wars with the player's role: none at region scale in the present frame. The chapter's conflict is a

private crime protected by a public system, and its resolution is exposure and rescue rather than a

campaign.

Section 9 — Minerals Resources and Trade Goods

Minerals and resources per resource, with Book-of-Trades entries: coral rag and reef limestone (the

Trade-8 building material, quarried, burnt for lime, and traded up the coast to reef-poor shores [SRC:

SRC_00196, p. 27]); mangrove timber and poles (structure, fuel, and an export in its own right [SRC:

SRC_00196, p. 19]); coconut fibre (cord, mats, rope, and hull stitching [SRC: SRC_00196, p. 7; SRC:

SRC_00198, p. 2]); sand, for the wells and the spits the ports sit on; and imported stone, which is the

region's most surprising resource line — igneous basalt and sandstone had to be brought to this

sedimentary coast, arriving as ships' ballast, and ballast stone was often reusable as building stone

[SRC: SRC_00198, p. 13; p. 12]. A town partly built out of the counterweights of the ships that made it

is a fact worth a whole quest.

Metals. The region's metallurgy in this shelf is monetary rather than martial: local production of

silver and copper coinage at the Kilwa mint, repeating names known from the chronicle [SRC: SRC_00166,

p. 12], with minting beginning in the middle of the eleventh century [SRC: SRC_00172, p. 1]. Iron is

attested for the coast in the geographers rather than in the Kilwa excavation — Al-Idrisi describes

iron-working settlements along the Sofala coast, often associated with gulfs or estuaries where ships

could enter [SRC: SRC_00196, p. 4]. Blacksmithing is therefore not a Kilwa home base and none is

asserted; the Trade-3 home is already rooted at South India.

Trade goods and routes. Gold is the spine of the system: Kilwa held the monopoly of the Sofala gold

trade from the twelfth century, the land of the Zanj supplied much gold to the Arab lands from the

tenth, and the fourteenth-century traveller records that Sofala's gold came from Yufi in the land of the

Limis, a month's march inland [SRC: SRC_00184, p. 3]. Ivory and, in later centuries, rubber and enslaved

people move on the same routes [SRC: SRC_00196, p. 8]. Inbound, the market is the whole ocean:

• Ceramics, the densest and best-dated import class. Persian hatched sgraffiato and Chinese stoneware

date the tenth- to thirteenth-century layers [SRC: SRC_00184, p. 29]; sgraffiato is the main Islamic

import to East Africa across the tenth to thirteenth centuries and is progressively displaced from

the later thirteenth by Chinese green-glazed stoneware [SRC: SRC_00181, p. 45]; from the fourteenth

century Southeast Asian wares enter the network, with Vietnamese underglaze-brown stoneware and

green-glazed Thai and Myanmar stoneware among the Songo Mnara finds [SRC: SRC_00181, p. 15]. Large

dishes and bowls with flattened rims over thirty-five centimetres across, in blue-and-white and in

Longquan green-glazed stoneware, became principal Chinese exports to the Indian Ocean [SRC: SRC_00181,

p. 25].

• Glass, beads, and the bead industry. Imported glass, glass beads and glassware travel with the

ceramics [SRC: SRC_00181, p. 29; SRC: SRC_00196, p. 24], and bead grinders — a class of artefact

recovered at many East African coastal sites, mostly of the late first millennium — show beads being

finished locally rather than only imported [SRC: SRC_00196, p. 10].

• Cloth, in both directions. Silk and cotton cloths and turbans on the wearer's back [SRC: SRC_00196,

p. 7], lengths of cloth as the price of an island [SRC: SRC_00184, p. 3 n8], and cloth as an offering

to the spirits of a place [SRC: SRC_00313, p. 3]. Cloth is this region's most socially loaded good and

the market should price it that way.

• Food and provisioning. Rice, wood, cotton and rope for boat repairs; water, food and firewood for a

port of call [SRC: SRC_00196, p. 25]. The unglamorous trade is the one a ship actually cannot sail

without, and it is the player's reliable early income.

What the goods MEAN is the section's L2 payload, and the record states it directly. Chinese ceramics

were not merely a display of wealth: they were consumed as tableware at public feasts, displayed in

intimate, honorific or sacred spaces, and worn as personal decoration, functioning as potent symbols of

social status [SRC: SRC_00181, p. 45]. Their value shows in how they were repaired — a late

fifteenth-century blue-and-white plate from Songo Mnara survives as five fragments pierced with rivet

holes so a metal wire could reconnect them, and a stoneware bowl base from Sanje ya Kati was carefully

cut into a new shape after breaking [SRC: SRC_00181, p. 19; p. 21]. And their density measures a

settlement's engagement with maritime trade — the ratio of imports at inland and rural sites is almost

insignificant beside a stonetown's [SRC: SRC_00181, p. 45; p. 29]. That gives the build a real economic

gradient: the same bowl means one thing in the harbour quarter and another an hour inland.

One reading the page carries as the record's own inference rather than as fact, because the record

frames it that way: the high proportion of small bowls in Chinese qingbai ware MAY be material evidence

of the domestication of rice culture and its related practices, and the medium and large tableware MIGHT

have been used in public ritual feasts alongside African and Islamic sgraffiato [SRC: SRC_00181, p. 45].

The hedges are the source's and the page keeps them. Related and equally hedged, and equally usable:

the fronts of mosques decorated with insets of imported bowls might be public statements by feast

patrons, suggestive of meals both eaten and yet to be consumed, with a Chinese porcelain inset recorded

on the front of a fifteenth-century mosque at Chwaka on Pemba [SRC: SRC_00181, p. 25]. A building that

advertises a man's generosity in porcelain is a social mechanic and a quest hook at once.

Currency and markets. The Kilwa mint's silver and copper, the sultan's tax on transiting goods, and

cloth as a store of value together give the region a three-layer money — coin, tax, and textile — and

the market beat at CH08_B02 is built on the difference between them. The reach is real and the page

states it without inflation: Kilwa coins of the twelfth century have been recovered in the Wessel

Islands off northern Australia, and the source itself raises the competing explanations for how they

got there rather than settling on one [SRC: SRC_00172, p. 1]. Carried as a genuine open question, which

is exactly the register the SQ12 rumour chain wants.

Section 10 — Music Sound and Composition Brief

Substrate-confidence flag, stated first because it governs this whole section: the music substrate for

this region is the thinnest lane on the page. The literal term census returns zero occurrences of ngoma

and zero of taarab across all 472 chunks. What the shelf attests about music is three sentences, and

the page will not build a musicology on them.

What is attested. At the Makunduchi New Year, as the burning hut is circled, women sing; in the evening

of the same day there are dances; and in the later twentieth century dance halls charging entrance fees

became an important source of festival income, while women and young men began singing obscene songs as

part of the public performances [SRC: SRC_00313, p. 3; p. 11]. Separately, the Indian Ocean world is

described as sharing to a great extent its architecture, cuisine, dress, music and literature [SRC:

SRC_00172, p. 11]. That is the whole of it: singing women circling a fire, evening dance, a modern

commercial dance-hall layer, and a general statement that the ocean's music was shared.

What the page builds from that, and no further. The region's audio identity is carried by its

soundscape rather than by an instrument roster the corpus does not supply: the call to prayer over

coral-stone streets, five times against the working day; the harbour — rigging, cordage, hulls working

against a quay, the market; the reef, which sounds different from open water and is the chapter's

danger cue; the monsoon itself as a low steady presence that changes direction between strata. The one

ensemble beat the page stages is the Section-18 festival window's, and it is staged at reference

register — women's voices circling a fire, drums implied and never sampled from a consecrated

performance, no ceremonial text authored or reproduced.

Composition briefs per track type, each cue carrying its adaptive-class tag {adaptive | loop |

oneshot} for Music Step 4 stem-splitting: an open-water crossing register for CH08_B01, wide and

patient and wind-timed, resolving into a landfall figure {adaptive}; a market-and-town bed for Kilwa,

many-voiced and busy, with the call to prayer cutting across it as a diegetic oneshot that the score

gives way to {loop, plus oneshot}; a reef register for the hazard zones, tight and arrhythmic and

listening {adaptive}; a warded-threshold cue at the reef climax, drum-led, held at reference register

and never reproducing a possession-rite performance {adaptive}; a grove register at the kaya edge, near

silence with the score stepping back for real forest, which is also the care posture {loop}; and a

naval-boss register across CH08_B09 to CH08_B11, escalating into the vril-charged third phase {adaptive}.

Sound design: harbour, market, reef, deck-and-rigging, mangrove creek and forest-margin terrain enums;

monsoon-sea, tide, gull and cordage ambient; the vril-cast tier for the aware-vessel confrontation and

the reef warding; inscription-activation at the Sealed-Letter read. Voice and audio tiering: bulk

ambient market and crowd lines route to the local no-clone tier; hero and named voices — the captain's

seamanship register, the scholar's reading register, the merchant's courtly Swahili-Arabic

code-switching — to the clone-authorized tier behind the §4.4.1 gate, with cross-cultural cloning

prohibited outright. The acquisition entry for a real Swahili music source stands in Section 16, and

until it lands this brief rests on thin substrate and says so.

Machine cue table (generated -- do NOT hand-edit)

Emitted by harness/emit_music_cues.py from the spine beat graph, this page's Zone Catalog, T0_SFX_Registry and T0_Theme_Registry. The composition briefs above stay prose because a composer reads prose; these rows are what the build consumes. Re-emit rather than edit.

Reading bed_source and licence_class together: sonniss_landed rows are direct use of a landed Sonniss file, and the Sonniss bundle licence prohibits use as training data -- no Sonniss file is ever a fine-tune corpus, a LoRA corpus, or an audio conditioning input for any generator. ace_step_worklist rows are generated beds (ACE-Step-1.5-XL-SFT, MIT) and must not be derived from any Sonniss row; each produced bed records its generating model and licence in T0_SFX_Registry.generation_licence_ref. composed_work rows are the composer-in-loop lane. The constraint travels on the row because this block is the surface that reaches canon.

layer_mode: base is the melodic statement a swap layer REPLACES -- layers are interchangeable, never additive, so the melody is present at every intensity. ambience_stack is the single declared exemption and it carries no melodic material.

Section 11 — Velheim Family Presence

Present at design grain, absent from every player surface. This is the chapter where the antagonist

apparatus first becomes directly visible in Africa through a single operative — NW_0007, held at Cat-5

and awareness FULLY_AWARE, the awareness-tier hinge of the whole arc. Per Josh's ruled Q1 handling the

family name is NOT announced, the pattern is NOT surfaced, and the player does not perceive an organized

family behind the man; Cassius Velheim does not appear and no institutional form, operative chain, or

thesis hint reaches any string, codex entry, journal line or environmental read.

What the player actually receives is exactly three things, and the page holds the ladder to them: a

recognition on sight that concealment cannot answer; the mechanism of that recognition, which is a

description circulating up the trade-lanes since the eastern seas and never a substrate read the early

arc could hold; and an unformed unease after the win that has no name attached to it. The dual win

route changes nothing about this — neither part-breaking the vessel nor exposing him hands over a

pattern.

The care line is the same one Section 4 states and it is not a stylistic preference: Abdallah's

corruption is his personal bargain, taken with full comprehension, and it is explicitly not a failing of

the honorific, of Islam, or of the Swahili community. The honoured sultan is a positive bearer of the

same elite and is present in the same chapter for exactly that reason. Combat routes through the man and

his own aware vril, never through the mosque, the orders, the lineage as a class, or the community.

Section 12 — All Characters Full Entries

The protagonist's path through the region references the spine (CH_08) rather than duplicating it: the

crossing and the apprenticeship, the landfall and the market, the trade home rooted, the reef traced to

Pemba and warded with the healers, the grove secured, the recognition, the dhow taken and the hold

faced, the letter read, the leaving. NPCs are MetaHuman-first; each entry carries the persona fields per

the entity-persona system.

The region's named cast (consumed from T0_Character_Index; personas FEED at the build pass):

• Sayyid Abdallah bin Rashid (CHAR_0055) — the merchant-prince and first fully-aware vessel; BE_0008;

Swahili-Arabic code-switching, courtly elegance, information gathered through conversation; the

chapter's one legitimate human combat target and the reason the care line above exists

• Nakhoda Binti Hassan (CHAR_0056) — the dhow captain, a rare female captain of the historical

tradition, the Trade-10 teacher and the first positive recognition; technical Swahili for the craft and

plain speech for the community; recurs at the Ch-9 and Ch-38 crossings while her ally status holds

• Kilwa Sultan al-Hasan ibn Sulaiman (CHAR_0057) — the documented historical ruler and Great Mosque

patron; formal Arabic at court and Swahili in private; never a combat target and never conflated with

the antagonist reading

• Ibn Battuta (CHAR_0058) — the documented fourteenth-century traveller, reached through the

time-fluctuation protocol; classical Arabic scholarly register; his own judgement of the city is

carried as the record gives it, that Kilwa was one of the most beautiful cities in the world from a

man who had seen Tangier, Alexandria, Cairo and Jerusalem [SRC: SRC_00006, p. 8]

• Scholar-Outsider Amara (CHAR_0059) — the wandering Sufi scholar, multilingual, the only NPC in the

chapter who can read the fairy-realm script and therefore the one who reads the Sealed Letter

• Mutota (CHAR_0060) — the interior gold-route trader; the Great Zimbabwe end of the web, reached

through the entry's era-composite framing

• Rasheed (CHAR_0061) — taken on the slave route and, on the good path, freed into the Sufi community's

keeping; a person with his own road toward refuge, never set-dressing and never spectacle

• Li Wen (CHAR_0062) — the porcelain merchant; the eastern end of the same web
• Ntala (CHAR_0002) — the healer's child, carried as active mystery with no on-screen beat here

Behind the named cast sit the region's documented offices — the roles the communities fill and refill.

Each is a role class rather than a minted character; the record documents the office, so none of these

proposes a name, and naming runs later under the naming method. Where an entry adds anything the record

does not carry, it says so. Every office below is benevolent or neutral; the region's malevolent

register stays where Section 5 puts it and no office is ever fused with it.

The mvyale — the descent-group religious expert (Zanzibar; the Section-18 window)

Within every descent group, religious experts called wavyale, singular mvyale, take care of the rituals

directed towards the local spirits; most are elders because expertise takes time to acquire, and both

men and women hold the office [SRC: SRC_00313, p. 3]. In the past the wavyale controlled access to land

through the rituals required for appropriating local resources, and their dues fell three times across

the millet cycle and again on hunting and on net and trap fishing [SRC: SRC_00313, p. 3; p. 3 n5]. At

the New Year those of them who lead the town-scale rites are called the elders of the year [SRC:

SRC_00313, p. 3].

• Personality: unhurried and exact about sequence; the office is a schedule as much as a power, and the

holder is the person who knows when, not the person who decides whether

• Background: authority held through a descent group and its inherited relation to a specific place;

contested rather than automatic, since positions of leadership were competed for and it was possible

to challenge a senior expert within one's own group [SRC: SRC_00313, p. 6]

• Own life and daily routine: authored from the office rather than the record — an ordinary

householder's round, interrupted by the calendar the spirits keep and by anyone who needs ground

• Relationship surface: the door to using land, water or a fishing ground legitimately; her regard

converts directly into where the player may work

• Three-path outcomes: ask before taking ground and carry the due yourself (generative — the town's

permission is real and it propagates); pay someone else to carry it (neutral); work the ground without

asking (corrosive — not a lightning bolt, but the standing of a person who takes without asking,

which on this coast is a durable thing). Non-gating throughout: refusal alters what the player may work

and who helps, never whether the chapter proceeds

• Dialogue hooks: the plain distinction between mila and dini stated without embarrassment; the flat

statement that a person can belong to more than one descent group and therefore owe more than one

place; the refusal to shorten a sequence

• Build notes: MetaHuman; protected practitioner, never a target; a Zanzibar office documented in

contemporary fieldwork, so it is seated in the Section-18 modern window and NOT relocated into the

fourteenth-century Kilwa present, and its material is never relabelled as general Swahili practice.

The source changed all its informants' names and the page carries none of them; the office enters, the

people do not. FEEDS T0_Character_Index as an unnamed persona row

The wa-ungwana householder — the stone-house patrician

Wa-ungwana designates the notables or patricians who occupy the large stone houses in the centre of the

towns [SRC: SRC_00184, p. 22 n93]. The record's own caution is the interesting half: a great stone house

proves a notable lived there and proves nothing about the town's weight in the world [SRC: SRC_00184,

p. 27]. The office is therefore built as status that must be read carefully rather than status that

announces itself correctly.

• Personality: courteous, exact about hospitality, and quietly attentive to what a stranger is worth;

status here is performed through what is set on the table and what hangs on the wall

• Background: a merchant household whose standing rests on an alliance between local authority and

foreign trade [SRC: SRC_00181, p. 45], which means his position is genuinely two-sided and genuinely

precarious

• Own life and daily routine: authored — the tide and the prayer times set his day; goods checked at the

quay, obligations at the mosque, a table kept for people who might matter

• Relationship surface: the region's most legible social ladder. Reading a house correctly — the porcelain

set in the wall, the size of the dish, whether the plate was repaired with rivets rather than replaced

[SRC: SRC_00181, p. 25; p. 19] — is a real investigation skill with real consequences at his table

• Three-path outcomes: earn a seat by useful work and read the room right (generative); trade with him and

keep it commercial (neutral); take advantage of his hospitality (corrosive — a house that has fed you

and been robbed for it talks, and this is a coast where reputation is the only passport)

• Dialogue hooks: the difference between what a house says and what a town is; the price of a bowl and the

price of the same bowl inland; the honest admission that his fortune depends on strangers

• Build notes: benevolent-to-neutral, never a target; the class includes Muslim householders of many

origins per Section 4's definitional anchor, and the build never renders wa-ungwana as a single ethnic

type. FEEDS T0_Character_Index, unnamed

The nakhoda — the master of a sewn ship

The office is the chapter's named ally's, and the page describes the class behind her rather than

duplicating her entry. Its knowledge is a real body of technique: the monsoon regime that decides the

sailing season; the pilot's arithmetic of daily runs with its own stated error terms — wind direction and

strength, currents, whether one dares sail at night, wreck hazards, and time lost to trade and repair

[SRC: SRC_00196, p. 4]; where a hull may touch and where the coral will open it [SRC: SRC_00198, p. 9];

and the stitched-plank craft itself [SRC: SRC_00198, p. 2].

• Personality: economical with words while working and generous with them afterwards; a person whose trade

punishes bravado

• Background: master of a vessel in a trade where the ship is the household; the record's own attestation

that this coast's ships carried votive engravings for their safety [SRC: SRC_00198, p. 2] says what the

relationship to a hull actually is

• Own life and daily routine: the wind's calendar, not the sun's — a season of sailing and a season of

waiting, refit and rope-work in between

• Relationship surface: the Trade-10 teacher relation, which is the arc's first returnable coastal home
• Three-path outcomes: carried by FORK_CH08_B at chapter scale — the promise kept, the promise left

ambiguous, the quiet parting, or the savings taken and a pre-dawn slip

• Dialogue hooks: what a day's sailing actually means and why the number is a lie in the wrong wind; the

engraving on the mosque wall and who cut it; why a stitched hull is better on this coast than a nailed one

• Build notes: benevolent; never a target; the class carries multiple genders on the historical record, and

the named captain is one instance of a real tradition rather than an exception invented for the chapter.

FEEDS T0_Character_Index, unnamed

The incorporated stranger — permission, service, and standing

The record carries a documented route from outsider to leader that is not a metaphor. A man came to

Makunduchi from the African mainland in the middle of the nineteenth century, received permission to

settle from the elders of the town, began in a subservient position, and rose to a leading role in the

town without any connection to a local mzimu at all [SRC: SRC_00313, p. 6]. Around it stand the terms

that make it real: authority in that field was both traditional and charismatic, positions were competed

for, and the field could be changed from inside [SRC: SRC_00313, p. 6].

• Personality: watchful, useful, and slow to claim anything; the man who rose did it by being needed
• Background: an office by achievement inside a system organized by descent — the exception the system

itself permits, which is the most interesting thing a system can have

• Own life and daily routine: authored — the work nobody else wants, done visibly and well, for years
• Relationship surface: this is the region's assimilation-integrity anchor and it is documented rather than

designed. Permission from the elders comes first, service second, standing third, and none of it is bought

• Three-path outcomes: ask leave and then earn the place (generative — the community's own attested path);

stay a guest and trade honestly (neutral); claim standing that was not granted (corrosive — on a coast

where a stranger's whole capital is permission, spending it is expensive)

• Dialogue hooks: the elders' grant, stated as a fact rather than a favour; what a subservient position

actually asks; the moment someone realizes the newcomer is now needed

• Build notes: a nineteenth-century Zanzibar case, so the CASE sits in the Section-18 window while the

PRINCIPLE — permission then service — is the region-general mechanic the chapter's assimilation lane runs

on. No real person is portrayed and the source's own name-changing is respected. FEEDS T0_Character_Index,

unnamed

The conservation mason — the modern-stratum office (Section 19)

The 2002-2004 works at Kilwa Kisiwani ran as a training site whose stated purpose was to train qualified

Tanzanian workers and transmit the knowledge of conserving coral-limestone ruins, alongside a cultural and

touristic revalorization of Kilwa within local development [SRC: SRC_00184, p. 11]. The works themselves

were surgical: remeshing masonry, replacing missing or damaged voussoirs, completely redoing the roof

waterproofing and the rainwater drainage, relaying floors, and shoring a collapsing fort wall [SRC:

SRC_00184, p. 11; p. 17]. There is also a documented reversal — two ungainly buttresses added in 1959 were

dismantled [SRC: SRC_00184, p. 16] — which is the office's whole ethic in one act: a previous generation's

fix, undone because a better understanding arrived.

• Personality: patient, technical, and unsentimental about ruins; proud in a specific and earned way
• Background: a craft learned on the monument itself, in a programme designed to leave the skill behind

rather than the scaffolding

• Own life and daily routine: seasons of work on a site, teaching as much as building
• Relationship surface: the Ch-59 revisit's keeper figure — the person who can tell the returning player

what has been held and what has been lost since

• Three-path outcomes: help the work and learn the technique (generative); pass through as a visitor

(neutral); take something from the site (corrosive, and this is the register in which looting a heritage

site should land — a named person's work undone)

• Dialogue hooks: why the buttresses came down; what lime does that cement does not; the difference between

a ruin and a building someone is still responsible for

• Build notes: benevolent; a modern office seated in Section 19 and never in the medieval present. FEEDS

T0_Character_Index, unnamed

Section 13 — Quest Framework, All Questline Types

Beat structure, fork shapes, SQ stages and gates, and spatial requirements COME FROM the spine's typed

graph (the §10, §10b and §11 rows and the §3b sites). The page adds region-scope quest dressing and

side-quest seeds. Everything below the main-story paragraph is proposal-tier region dressing consumed

from this shelf: each row is cited, each is opt-in, each is non-gating. No proposed_quest_type is asserted

here, per schema v0.2.2 §11.1.

Main story (consumed, not re-authored): the crossing and the apprenticeship; the market as a language; the

Trade-10 home rooted; the reef traced to Pemba and the healers' line between the crafts learned; the safe

lane read and the warded threshold held; the grove secured and the elder's recognition earned; the

recognition on sight; the boarding, the hold, and the aware-vessel confrontation; the letter read; the

leaving. Two forks carry the durable divergence, the hold (FORK_CH08_A) at elevated care and the leaving

(FORK_CH08_B).

Region-scope seeds, each grounded and each non-gating:

• The votive engraving. Ship engravings occur on the outer walls, outer windows and door jambs of mosques

and near the main entrances of domestic buildings, and are read as charms or votive offerings for the

safety of the vessel depicted [SRC: SRC_00198, p. 2]. The chain: find them (they are genuinely easy to

walk past), learn what they are, and at the Trade-10 tier where it is earned, have one cut for a ship

the player cares about. The Songo Mnara engraving with its swan-necked bow and pennant and the three

Great Mosque engravings with their oculi and tasselled amulets are the reference set [SRC: SRC_00198,

p. 3]. Reward: a lasting mark on a vessel, and a lore chain that ends at the Trade-10 tier gate.

• Ballast and the built town. Basalt and sandstone are not native to this sedimentary coast and arrive as

ships' ballast, which is often reusable as building stone [SRC: SRC_00198, p. 13; p. 12]. A stone in a

wall can therefore be traced to an ocean. The chain: notice the wrong stone, follow it to the beach

concentrations at Matuso and Kivurugu, and read whether a scatter is a wreck or a deliberate unloading

— which is exactly the judgement the archaeologists themselves have to make [SRC: SRC_00198, p. 13].

• The harbour wreck. In the Kilwa Kisiwani port anchorage a stone anchor lies near a concentration of

ballast and thirteenth- to sixteenth-century pottery, including light incense burners, bones and

complete bowls that would not be normal ballast choices — which is why the excavators prefer a

shipwreck to a dumping [SRC: SRC_00198, p. 12]. A diving-and-inference beat whose correct answer is the

historical one, and whose reward is a salvage the player must then decide what to do with.

• Jiwe la Jahazi. The island's own oral tradition tells of a wreck at the sharp coral-limestone crest,

and the finds there — late eighth- to tenth-century pottery of types also found at Siraf and Sohar,

with basalt ballast — support a Gulf vessel lost on the reef [SRC: SRC_00198, p. 9]. The story comes

first, the evidence second, and neither cancels the other. Reward: the SQ1 register applied to a real

place, and a hazard the player can now read.

• The riveted plate. A single late fifteenth-century blue-and-white plate survives in five fragments with

rivet holes for wire repair [SRC: SRC_00181, p. 19]. Reassembling a broken import, or carrying one to

someone who can, is a small craft loop that teaches the region's economics better than a ledger would.

• The tide window at Serenge. The ford onto the island works at spring tides and comes out near the

Makutani enclosure [SRC: SRC_00184, p. 10]. A timing route with a real consequence for being late,

usable both as a shortcut and as the reason a pursuit can be broken.

• The lime burn. Four superimposed phases of lime kilns in one sequence, calibrated 1022-1329 [SRC:

SRC_00166, p. 13], and lime and reef coral traded to shores that had none [SRC: SRC_00196, p. 27]. A

Trade-8 production loop with a genuine supply chain and a genuine export.

• The provisioning run. A port of call exists to give ships water, food and firewood and to repair them

[SRC: SRC_00196, p. 25]. The least glamorous and most reliable early work on the coast, and the honest

way a child earns a bench.

• The Bangwe stone. A ritual site linked to the legend of a petrified imam who did not comply with

Koranic rules, at a site the population highly respects and visits to ask for healing and protection

[SRC: SRC_00006, p. 77]. A discoverable site with a story, staged with the community's own respect

intact — the legend is theirs and the player is a guest at it.

• The coin question. Kilwa copper coins of the twelfth century recovered in the Wessel Islands, with the

source itself weighing competing explanations [SRC: SRC_00172, p. 1]. An SQ12 rumour chain whose honest

ending is a question, which is a legitimate ending.

• The empty houses of Songo Mnara. Exceptionally well-preserved fifteenth-century buildings on a site that

the imported-ceramic count says was never really in the international trade — a minor satellite, perhaps

founded by refugees from an older settlement, whose stone says one thing and whose rubbish says another

[SRC: SRC_00184, p. 27]. The best investigation premise the shelf contains: a place that looks more

important than it was, and the reason why.

• The mangrove round. Poles cut and carried to the landing place for export, whelks taken from the same

forest, traps built where the boats come in [SRC: SRC_00196, p. 19; p. 13]. A gather circuit that is

also a map lesson, because the creek routes are how a person learns this coast.

Every seed above is opt-in and none gates the main line. Where a seed implies a new registry row, a new

zone or a new named character, it is carried as a FEED and asserted nowhere else.

Section 14 — Trade Skills, Detailed Content

Home-base trade content: TRADE_10 Fishing and Navigation roots its ANCIENT home base here, paired with a

modern home at Ch 56 — the protagonist's first returnable coastal rootedness and the first bench that is

theirs to come back to. The teacher is Nakhoda Binti Hassan and the curriculum is the coast's own.

Trade-10 curriculum, grounded:

• The sewn hull. The craft's defining technique is stitching rather than nailing, and the shelf's

attestation of it is Ibn Jubayr's record of vessels at Aydhab, carvel-built with planks stitched with

coconut fibre through holes bored at intervals near the edges of adjacent planks [SRC: SRC_00198, p. 2].

A comparable thirteenth- to fifteenth-century hull excavated in Kerala shows the related family of

solutions — double planking, cleats carved into the inner planks with rope fragments still in them

suggesting the planks were held in compression by lashings, and shallow mast-steps only fifty

millimetres deep that must have needed further support [SRC: SRC_00198, p. 4]. The player learns holes,

cord, tension, and where a stitched hull's strength actually lives.

• Why stitched. The record gives the reason a stitched hull suits this coast: hulls of the period were

designed to flex going over shallow reefs and sand bars [SRC: SRC_00198, p. 9]. A boat that gives is a

boat that survives ground a rigid one would not. That is the single best worldview lesson in the

region's trade, and it is literally true.

• The rig. The coast's own images show yards supporting square sails with a spreader yard at the foot of

the sail, on hulls with curved flared bows [SRC: SRC_00198, p. 3]. The rigging traverse — the chapter's

agility course — is built on that geometry rather than on a generic lateen.

• Ground tackle. Three stone anchors were recovered at the channel entrance to the old port of Mombasa,

one an Indo-Arabian type and two composite — wooden anchors weighted with stones; an Indo-Arabian shank

1.2 metres long and 0.35 wide was recorded at Kilwa [SRC: SRC_00198, p. 3]. Two anchor types with

different behaviours is a real choice with a real cost.

• Pilotage arithmetic. One day's sailing as roughly eighty to a hundred kilometres, with the error terms

the record itself names — wind direction and strength, currents, night sailing, wreck avoidance, and

time spent trading and servicing [SRC: SRC_00196, p. 4]. The navigation minigame is that estimate and

those corrections, and the reward for mastery is a smaller error bar rather than a bigger number.

• Reading a landing place. Where a vessel can touch is knowledge: sand spits and inlet mouths with deep

channels for larger ships, versus mangrove passages usable only at high tide, versus the shallow creeks

where only a small narrow craft works [SRC: SRC_00196, p. 17; p. 25; p. 13]. The port taxonomy the

record sets out — entrepôt, lesser commercial port, port of call — is the trade's own map legend [SRC:

SRC_00196, p. 25].

• The small craft. Beside the ocean-going vessel the record describes a narrow local boat, carvel-planked

and narrowing at bow and stern, shaped more like a canoe than a dhow, that can be paddled or rowed,

carries a lateen sail, takes its rudder off the side of the stern, and is useful for navigating many

shallow inlets [SRC: SRC_00196, p. 13]. The page carries it as the class of small boat this coast's

creeks require without asserting the fourteenth-century name of it, since the record's own name for the

vessel it describes is a modern Angoche one.

• The gather side. Fish traps built at the landing place; whelk, conch, clam, cockle and olive shell from

four distinct grounds; mangrove poles cut for export [SRC: SRC_00196, p. 13; p. 19]. Trade-10's gathering

loop is the coast's real subsistence, and it feeds the provisioning economy rather than a crafting bench

in a vacuum.

The vehicle line. The mtepe anchors here as the arc's first ownable vessel (VH_004; the registry row is a

POPULATE target, not a page assertion) — crewed first under the captain, then earned or commissioned as a

hull of one's own. It is strictly non-gating: a travel convenience and a returnable home on the water,

never a wall, with the crossing to Ch 9 and every boss and quest line fully playable without it. The

build honours the craft as a real documented naval-engineering achievement of this coast, and the honest

provenance is stated in Section 5's naming block: the coir-and-no-nails formulation is spine and RCM

canon, while this shelf's own attestation is the Aydhab stitching record and the coast's own engravings.

Other trades, at seed level: Trade-8 Engineering takes the lime, kokoto and causeway seeds from Section 7;

Trade-6 Tailoring takes the cloth economy of Section 9 as substrate without a home base; Trade-4 Herbalism

takes nothing from this shelf, because the shelf attests no medicinal botany for this coast at all, and

Section 16 records that rather than importing a neighbouring region's pharmacopoeia. The obligation

economy that gives trade output somewhere to go is the region's own attested one and it is documented

rather than designed: ritual dues owed three times across the millet cycle and again on hunting and on

net and trap fishing, and permission to use ground obtained through the same channel [SRC: SRC_00313,

p. 3; p. 3 n5]. What the player gathers, catches and makes is what pays the dues that keep a stranger's

standing good — tying the Trade-10 round to ws.community_state_per_region, ws.identity_hiding_status and

the assimilation-integrity mechanic, with the modern-stratum caveat Section 12 states, and with

progression never touched.

The vocabulary this section and Section 5 introduce — mila, dini, uganga, uchawi, mzimu and mizimu,

mvyale and wavyale, wa-ungwana, nakhoda (spine-canon consumed, unattested in this shelf), kokoto, mchele,

nguo nyingi — takes Section-17 glossary entries

so localization does not flatten it. Words the page deliberately does NOT add to the glossary on its own

authority are listed in Section 5's naming block, which also declares the two consumed terms that carry

no shelf attestation behind them.

Section 15 — Cross-Regional Connections

Connections to visited regions, serpent-thread continuity, and vital-force parallels. Every connection

names its Cross_Cultural_Link_Graph row, and each bullet characterizes exactly one link so that no type

claim spills onto a neighbour. Reciprocation discipline binds: a page-side claim without a reciprocal

target row is a defect.

• Sri Lanka to the Swahili Coast, the vital-force term: LINK_0002, registry link_type

parallel-independent — the third independent appearance of the tendi parallel after the Batak and Vedda

legs, a convergence across a whole ocean and never a diffusion claim. The Swahili end of that edge is

the one this page flags: Section 5's naming block records that the word occurs zero times in the

region-filtered shelf, so the page consumes the thread-canon assignment and mints no gloss of its own

• Sri Lanka to the Swahili Coast, the ill-intentioned spirit: LINK_0149, registry link_type

parallel-independent — the yakku and the coast's own malevolent sea-spirit as two independently reached

answers to the same fear, each named and warded by its own tradition. The Swahili end is grounded in

the shetani-and-mizimu material of Section 5 and in the spine's Mackenrodt and Giles anchors, which are

not in this shelf

• Bali to the Swahili Coast, the two branches of one craft: LINK_0168, registry link_type

parallel-independent — pengiwa and penengen beside uchawi and uganga. This shelf gives the Swahili end

its strongest possible attestation: harming and healing as two branches of one craft, using the same

means to opposite ends, with the boundary between them acknowledged as ambivalent by the tradition

itself [SRC: SRC_00313, p. 2]. The point of the edge is that each tradition draws the line internally,

never that an outside morality was imported into either

• The Swahili Coast toward the Maya cenote: LINK_0009, registry link_type catalog-membership — the coral

reef entering the geological vril-amplification catalog whose next member is the Ch-48 karst. A

catalogue of natural formations, never a lost-engineering claim, and the reef is deliberately NOT

claimed for the man-made anomaly class

• The Swahili Coast toward Persia: LINK_0064, registry link_type sets-up — the Sufi chain opening at

Kilwa and deepening at Ch 19

• The Swahili Coast toward Moorish Iberia: LINK_0065, registry link_type sets-up — the same chain's

western reach at Ch 31

• Moorish Iberia to the Swahili Coast, the manuscript circuit: LINK_0179, registry link_type

knowledge-transmission — the Indian-Ocean Islamic book trade arriving here, with the guardrail the

registry itself carries, that this is a shared scholarly network in which every node is an author. This

shelf's own support for the receiving end is the madrasa-and-literacy record and the observation that

Arabic was the common literary and commercial language of the whole ocean [SRC: SRC_00172, p. 8; p. 4]

Beyond the rowed edges, four connections this shelf opens and the page carries as OWED rather than done,

because minting a Cross_Cultural_Link_Graph row is a registry action and several target ends have no page

here to receive them:

• The Swahili Coast and Great Zimbabwe. Sofala, Kilwa's vassal, traded with the kingdom of Great

Zimbabwe, whose prosperity the gold profits materially served and which worked many mines on the high

plateaus of Sofala's hinterland [SRC: SRC_00184, p. 3 n13]. The interior end is a real polity with its

own agency and its own mines, and the edge is an economic relationship between two states rather than a

coast civilizing an interior. Reciprocal row owed at the Ch-13 Great Zimbabwe treatment

• The Swahili Coast and the Hadramaut. The Mahdali dynasty that ruled Kilwa from 1277 came out of the

Hadramaut [SRC: SRC_00184, p. 3], and south Arabian contact shows most specifically in the Swahili

vocabulary of ships and shipbuilding, among the northern Swahili closest to that coast and most mixed

with Hadrami people [SRC: SRC_00172, p. 8]. An attested movement of people, rule and technical

vocabulary in both directions, owed to the Arabian pages when the arc reaches them

• The Swahili Coast and China. Chinese ceramics reach far inland and an African giraffe went east as a

gift to the emperor [SRC: SRC_00006, p. 8]; Chinese green-glazed stoneware, blue-and-white porcelain and

the specialty large dishes made for the Indian Ocean market arrive in quantity [SRC: SRC_00181, p. 45;

p. 25]. Both directions attested in one relationship, owed to the China pages

• The Swahili Coast and Southeast Asia. From the fourteenth century Vietnamese, Thai and Myanmar

stonewares enter the same network and turn up at Songo Mnara [SRC: SRC_00181, p. 15]. The arc has

already walked the other end of that route in Ch 4-6, which makes this the quiet payoff of the first

three regions: the player has stood where these bowls were made

Serpent-thread continuity: Thread 2 anchors here through the interior trader carrying the Great Zimbabwe

python tradition, reached through the entry's era-composite framing and treated fully at Ch 13; this page

asserts no serpent content of its own, because this shelf carries none. The arc's larger corruption

pattern is held dark at page grain per Section 11.

Section 16 — Academic References and Books to Acquire

Every citation carries its use_class (SUBSTRATE · ESOTERIC-PRIMARY · AUTHORIAL-REBUTTAL · REFERENCE ·

TONE · DUPLICATE · FRINGE-SKIP). Book-substrate citations arrive via the query layer with SRC_ ids; the

retrieval-provenance trail rides the page. These are the region-FILTERED hits (query_substrate --region

swahili_coast, 472 chunks across 8 sources), and the filter matters here for the same reason the pilot

page records: an unfiltered query for another region's architecture returned Kilwa, so region-scoping is

a correctness requirement rather than a convenience.

Priority books and processed substrate:

• Kilwa dans Annales islamologiques (Pradines and Blanchard, "Kilwa al-Mulûk") — the keystone. The

Sections 1, 5, 6, 7, 8 and 9 anchor for the architecture, the dynastic sequence, the fortifications, the

funerary record and the conservation programme; 94 chunks, French [SRC: SRC_00184] · SUBSTRATE

• Global Trade and Swahili Cosmopolitan Material Culture (Zhao) — the Section-9 trade-goods and

social-meaning anchor; the strongest single source on what imports MEANT [SRC: SRC_00181] · SUBSTRATE

• Settlement and Trade from AD 500 to 1800 (Pollard et al., Angoche survey) — the Section-1, 2, 3 and 14

landscape, subsistence and port-taxonomy anchor. NOTE the scope caveat this page applies throughout: its

fieldwork is Angoche in northern Mozambique, not Kilwa, and its crop list is explicitly modern, so its

Kilwa material is carried where it cites Kilwa and its Angoche material is attributed to Angoche

[SRC: SRC_00196] · SUBSTRATE

• Shipwreck Evidence from Kilwa, Tanzania (Pollard, Bates, Ichumbaki and Bita) — the Section-1, 13 and 14

maritime anchor: the reef, the causeways, the wrecks, the anchors, the ship engravings

[SRC: SRC_00198] · SUBSTRATE

• Dating Kilwa Kisiwani: a thousand years (Wynne-Jones, Horton, Fleisher and Olsen) — the Section-18

chronology anchor and the corrective on the immigrant-founding hypothesis [SRC: SRC_00166] · SUBSTRATE

• Early Globalization in the Indian Ocean World: An East African Perspective (Lodhi) — the Section-4

language, literacy and Indian-Ocean-integration anchor; a synthesis rather than a field report, and used

for its linguistic and integration argument [SRC: SRC_00172] · SUBSTRATE

• The Makunduchi New Year study (Echtler, "They Bewitched the Generator") — the Section-5 spirit-cosmology

and Section-4 festival anchor, and the ONLY processed source attesting the uganga and uchawi distinction.

Scope caveat applied throughout: Makunduchi in southern Unguja, contemporary fieldwork, seated in the

Section-18 modern window and never relabelled as general or medieval Swahili practice

[SRC: SRC_00313] · SUBSTRATE

• Kilwa Inventory Catalogue (Kilwa Islands Tour Guide Association, EU-funded, 2014-2016) — a heritage

inventory and tour-guide document, heavily degraded in OCR. Used for its ritual-sites statement, the

Bangwe stone legend, and its framing of Kilwa's reach and its scars; it is a community heritage

publication rather than a peer-reviewed study and the page cites it only where its statements are clean

[SRC: SRC_00006] · REFERENCE

MUST-ACQUIRE (specialist primary anchors not yet in the corpus; the four highest-value gaps first):

• A Mijikenda kaya source. This is the page's largest hole and it is a §17.1 hole, not a convenience one:

a literal census across all 472 chunks returns ZERO occurrences of Mijikenda, kaya or Giriama, and the

semantic query returns only unrelated Kilwa and Angoche pages. Thread 5 anchors on the kaya forests, the

chapter stages a sub-boss at their edge, and the page can currently say nothing about them beyond the

spine's own UNESCO-inscription line. Until a source lands, the build authors a forest margin that

identifies no real grove and reproduces no sacred site.

• Mackenrodt, Swahili Spirit Possession and Islamic Healing in Contemporary Tanzania, and Giles on Swahili

spirit cults. These are the spine's own named anchors for the sea-jini cosmology, the waganga and the

warding rite, and neither is in the processed corpus. Acquiring them closes Section 5's naming block.

• Horton and Middleton, The Swahili: The Social Landscape of a Mercantile Society; Middleton, The World of

the Swahili. The RCM lists both and the corpus has neither. They are the anchors for everything Section 4

currently reports as an honest gap — dress, household, gender, kinship, the texture of ordinary life.

• A Swahili music source. The census returns zero ngoma and zero taarab, so Section 10 rests on three

sentences and says so.

Also owed, at lower priority: Pouwels, Horn and Crescent, for the religious history that would let Section

5 seat the Sufi orders on its own authority; Sutton, Kilwa: A History; a fourteenth-century Kilwa

archaeobotany, which exists (the 2016 excavation floated every context) and would close Section 2; and a

Swahili arms-and-seamanship source for Section 6.

Verify list, carried forward: the tendi attestation (zero occurrences, Section 5); the waganga and

ngoma-ya-majini term check (zero occurrences, Section 5); the Sufi order names at Kilwa (zero occurrences

of both order names); the Ajami script attestation (zero occurrences); the nakhoda term check (zero

occurrences, consumed from CHAR_0056 rather than supplied by this page); the mchawi term check (zero

occurrences, consumed from the BE_0008_S1 registry boss_name), both of these last two against the same

Mackenrodt and Giles acquisition; the mtepe coir-and-no-nails

formulation, which is RCM canon and not this shelf's; the fourteenth-century Kilwa costume; and the

traveller's visit date, which the RCM and one source give as 1331 and another as 1332 [SRC: SRC_00172,

p. 4] — recorded so a later reader does not silently correct one to the other.

Rabbit holes (minimum five): the reef-coral causeway as deliberate navigational infrastructure; the

keystone-free counter-thrust arch in coral rag; the ballast-to-building-stone chain and what it implies

about where a town's walls come from; the Kilwa mint and the Wessel Islands coins; the porcelain-inset

mosque façade as a public statement of feasting; and the archaeology of the twelfth-century dietary shift

away from fish.

Section 17 — Unreal Engine 5.8 Build Specifications

Environment areas, interiors, vegetation palette, NPC population and behaviour, performance budget —

re-pointed to the researched UE-5.8 chain. Registry data lands as DataTables (one C++ USTRUCT per

registry: the Zone Catalog, the boss thin-index, the creature roster rows). PCG consumes the Zone Catalog

plus DataTables to scatter the mangrove creek, the plantation belt, the reef flat and the forest margin.

The Landscape import runs the Python path over the Section-1 DEM tiles. Scene rows drive Sequencer. The

three-seam routing governs which agent surface executes which step. MetaHuman batch generation

round-trips Epic's cloud for the named cast, and the crowd must carry the Section-4 definitional anchor

into its generation parameters: a Swahili port crowd is many origins holding one culture, so a single

appearance archetype for the market population is a §17.1 defect, not a budget saving.

Water and vertical are this region's two build risks and both are named early. Three of five zones are

water or water-edge, the boss set-piece is a moving deck on open sea with a rigging traverse above it, and

the reef puzzle is read at sea level while the hazard lives below it — so the water rendering, buoyancy and

moving-platform work is a region-specific dependency rather than a shared cost, and it is flagged to the

build plan here. The Husuni promontory's rock-cut sea stair and the harbour quay are the two vertical

set-pieces and both are real documented geometry (Section 7).

Registry dependency, recorded as a gate rather than a note: the T0_Region_Index row for swahili_coast

carries region_id, region_name, chapters, modern_country, coordinates, time_period, page_status,

last_updated and node_kind, and NOTHING else — node_class, the environment-profile pack, the movement and

terrain columns, and all three SFX id arrays are empty where the sibling region rows are populated. The

page consumes what the row has and asserts nothing for what it lacks; the row is the registry pass's to

fill, and the build cannot ship this region until it is.

Localization-ready discipline binds from the first page: every player-facing string lands in string

tables (FText and String Table flow), never hard-coded — including the Kiswahili and Arabic terms this

page introduces (mila, dini, uganga, uchawi, mzimu and mizimu, mvyale and wavyale, wa-ungwana, nakhoda —

which is spine-canon consumed and unattested in this shelf, per Section 5's naming block —

kokoto, mchele, nguo nyingi), each carrying a glossary entry with a meaning-carrying translation rather

than an opaque transliteration: the man of many clothes for nguo nyingi, uncooked rice for mchele, the

hammered coral gravel for kokoto, the notables of the stone houses for wa-ungwana. Husuni kubwa itself

carries its gloss, the great fort, which the record states outright [SRC: SRC_00184, p. 14 n64], and the

page surfaces it on first introduction so the palace reads as a named place rather than a label. Terms the

page refuses to add on its own authority are listed in Section 5's naming block, and the call to prayer,

the Koran recitation at the mizimu, and any ceremonial text are referenced and never authored.

Performance budget: platform-tiered frame-rate and triangle and draw-call budgets; LODs on the coral-rag

façades, the mangrove foliage and the rigging; World Partition streaming across the archipelago with the

open-sea arena as its own streamed volume; and the protected sites — Kilwa Kisiwani, Songo Mnara, the

kaya forests — built as type-accurate referenced set-dressing rather than high-poly hero reproductions of

the actual monuments.

Section 18 — Time Period and Fluctuation Design

Time display and fluctuation events: this region's strata are unusually well dated, because a

single-context excavation with a full radiocarbon programme sits under them [SRC: SRC_00166, p. 9]. The

strata the build supports, each with what changes on screen:

• c. 800-1000. Early Gulf ceramics — white glazed and turquoise glazed wares — and TIMBER buildings; grey

sandy layers with fish bone, ash spreads and small pits, then multiple floors, daub walls and post holes

across at least six buildings [SRC: SRC_00166, p. 13]. There is no stone town yet. A player who reaches

this window finds the same place made of wood and thatch, which is the single most effective way to

teach that the coral city was built by the people who were already there.

• c. 1000-1200. Sgraffiato pottery, timber and daub architecture, the first written mention of Kilwa around

1060, and a letter recording an Ibadi presence around 1115 [SRC: SRC_00166, p. 13]. A tower is built.

Also the window of the Kilwa-Shâgh wars and of walled Sanje ya Kati [SRC: SRC_00184, p. 3; p. 29].

• c. 1200-1290. Later sgraffiato, and the two things that arrive together — the earliest coins and the

earliest stone buildings [SRC: SRC_00166, p. 13]. The lime kilns run through this band, calibrated

1022-1329. This is where the city the player knows starts becoming itself.

• c. 1290-1400, THE PRESENT FRAME. Mahdali coins, the Husunis, the Great Mosque extension, Husuni modelled

wares and celadons; a plaster floor of about 1300; the Great House built; an enclosure wall raised in

porites blocks; and the traveller's visit in 1331 [SRC: SRC_00166, p. 13]. The palace is dated by its own

dedication to a sultan reigning 1320-1333 [SRC: SRC_00184, p. 14].

• 1466-1476. Succession civil wars, and vassals such as Sofala and Zanzibar emancipating themselves [SRC:

SRC_00184, p. 3]. The hegemony coming apart.

• The Portuguese window, 1498-1512 and after. The Portuguese fort raised 1505 and dismantled 1512, the

sultanate nearly destroyed in 1505, and the 1519 grounding off Kilwa [SRC: SRC_00184, p. 17; SRC:

SRC_00172, p. 1; SRC: SRC_00198, p. 26]. Songo Mnara is abandoned in the second half of the sixteenth

century [SRC: SRC_00184, p. 27].

• The Omani and slave-trade window, eighteenth and early nineteenth century. The Makutani enclosure's

essential role in the eighteenth-century slave trade and the mainland road that fed it through the

Serenge ford [SRC: SRC_00184, p. 10]; Songo Mnara briefly reoccupied by a few families [SRC: SRC_00184,

p. 27]; the Gereza fort founded 2 April 1807 on the Portuguese ruin [SRC: SRC_00184, p. 17]. This window

carries elevated care and is the honest historical weight of the region, routed through the trade's

apparatus and never through the community.

• The modern window. The Makunduchi festival material of Section 4 and the wavyale material of Section 5

both belong here rather than to the medieval present, along with the twentieth-century political

reshaping of the festival that the source traces [SRC: SRC_00313, p. 11], and the 2002-2004 conservation

programme [SRC: SRC_00184, p. 11].

Switching mechanics: per the Geography §11 time-fluctuation model, this is still an early node, so

switching is site-cued rather than free — the cue sites are the Great Mosque (whose own fabric is layered

across the strata), the Husuni promontory, the Serenge ford, and the reef crest, each of which physically

changes between windows. Era glimpses: the timber town under the stone one; the coral city at its height;

the fort inside the fort; the quiet village after.

Section 19 — World State Versions, Modern Revisit

After Vimana activation (Ch 55) the player revisits the world from the air and sees the integrity-mapped

world states. This region's revisit is the Ch-59 Africa continental-revisit, which is the coherent modern

consumer of the Swahili-coast marks — the leaving mark (FORK_CH08_B) read for the earned-standing

keeper-stands-with and known-guest receptions, and the Sufi community's standing from the hold

(FORK_CH08_A) riding alongside it.

The DERIVED texture bands are {Truly Good · Good Enough · Drifting · Evil · Pure Evil (terminal)}; the

world-state NAMES are {Luminous · Restored · Neutral · Damaged · Corrupted} (Neutral exists only as the

WS3 world-state name, never a band or an ending label); endings resolve by the Ch-65-locked path_bucket

{TRULY_GOOD, GOOD_ENOUGH, EVIL}. At Ch 8 the region initializes at integrity band L3 Drifting with

world-state WS3 Neutral and path_bucket UNLOCKED, the lock being far ahead at Ch 65.

What the revisit actually shows, and it is the region's most affecting design fact: the city the player

knew is a ruin, and the ruin is cared for. The record is explicit that the city eventually turned into a

quiet village and that the remaining architectural sites are inscribed on the World Heritage List [SRC:

SRC_00006, p. 8], and that a conservation programme ran on the Great Mosque, three tombs of the Malindi

cemetery, the domed mosque, the palace and the fort between September 2002 and December 2004 [SRC:

SRC_00184, p. 11]. Visual differences per state at the Ch-59 read are therefore not about the monuments

falling down — they fell down centuries ago — but about the community around them: whether the player is

recognized as someone the coast remembers well, and whether the conservation mason of Section 12 is glad

to see them. A generative leaving reads as recognition and aid and an open door at the training site; an

ambiguous one as a cooler courtesy; a corrosive one as a coast that remembers a theft and receives the

returning player accordingly. Non-gating throughout: recognition opens or alters aid and lore, never a

wall. Modern Velheim presence per state: none surfaced at this region, which is reveal-ladder-safe at

Ch 8's position and remains the page's posture at the revisit.

Section 20 — Session Planning and Source Guide

Per-region population plan (which staged SRC_ sources feed which sessions): the Section-1 and Section-7

geography-and-architecture session draws SRC_00184 with SRC_00198 for the maritime landscape; the

Section-9 trade session draws SRC_00181 with SRC_00166 for the coinage and SRC_00196 for the port

taxonomy; the Section-4 people-and-language session draws SRC_00172 and SRC_00166 with SRC_00184's

definitional footnote as its anchor; the Section-5 spiritual session draws SRC_00313 with SRC_00006's

ritual-sites statement, and is the session most dependent on the MUST-ACQUIRE Mackenrodt and Giles; the

Section-2 and Section-3 flora-and-fauna session draws SRC_00196 and SRC_00166 and is thin by nature; the

Section-14 trade session draws SRC_00198 and SRC_00196; and the Section-18 chronology session draws

SRC_00166 and SRC_00184.

Processing status against the book pipeline (inbox → processed → registry): all eight cited sources are

processed and query_ready in book_substrate with region_tags swahili_coast, 472 chunks total, and the

retrieval is region-filtered per the substrate-region ruling. Two of the eight are non-English or heavily

degraded (the keystone is French; the heritage catalogue's OCR is badly broken), which is why several of

this page's claims are carried at a coarser grain than the pilot page's equivalents.

Build-ready date: the machine layer (Zone Catalog, Derived Indexes) is build-ready now against the v3.1

spine, with two blockers that are NOT page work and are recorded here so they are not lost. First, the

T0_Region_Index row's empty environment-profile column pack (Section 17). Second, the Mijikenda kaya

absence (Section 16), which blocks not the build but the AUTHENTICITY of one of five zones and one of

three bosses. The Section-4, Section-6 and Section-10 depth is gated on the Horton-and-Middleton,

arms-and-seamanship, and music acquisitions respectively. Everything else on this page stands on the

shelf it cites.

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