music/ORNAMENT_AND_MELODIC_SURFACE.md
CANON SUBORDINATION — this document is PROPOSAL-TIER: it serves canon and never outranks it.
Canon served: the round-2 grading floor at the tail of docs/spine/DECISIONS_PENDING_JOSH.md
(grep "ROUND 2 GRADED"), the era column of registries/T0_Chapter_Index [ACTIVE v1.2] for
CH_PROLOGUE through CH_13, Section 10 of every slice page under _source/02_Tier_2_Region_Pages/,
and the care postures those pages state in their own words.
If this document disagrees with canon, CANON WINS and this document is the defect.
TIER: RESEARCH SYNTHESIS, PROPOSAL-TIER. Canon is authority; nothing here re-decides a region's
reading, its era, its instrumentation or its care register. Every substantive claim carries a source —
a canon file:line for our own decisions, a URL or a named work for the world's. Where the record is
thin, contested or absent this document says so in the sentence rather than in a footnote, and it
never presents a reconstruction as an attestation.
The ROUND 2 GRADED entry at docs/spine/DECISIONS_PENDING_JOSH.md:5600 and following, verbatim:
"Neither of the songs pass. They sounded much better, but the melodies are way too simple. There
needs to be a lot more complexity added and to put some intelligence into the melodies and how
everything orchestrates... Remember all of the culture and regions we are gonna be working with from
ancient history, mythological, and modern day as well up to even technological and techno music... my
words and reviews are just the floor, and you are required to enhance and build within our vision."
Then registries/T0_Chapter_Index [ACTIVE v1.2] for the fourteen node eras; Section 10 of
flores_island.md, bali.md, sumatra_java.md, sri_lanka.md, south_india.md,
swahili_coast.md, ethiopia.md, rift_valley.md, central_africa_congo.md, west_africa.md and
south_africa.md; each page's represent-with-care paragraph; docs/spine/CH_PROLOGUE.md and
docs/spine/CH_01.md music entries; the sibling lane's
docs/proposals/music/craft_research/ORGANOLOGY_SLICE_CULTURES.md; and the round-2 composer itself at
harness/music_gen/pass2_flores.py and harness/music_gen/pass2_realise.py.
Four siblings own the neighbouring territory and this lane cites rather than repeats them.
ORGANOLOGY_SLICE_CULTURES.md says what instruments exist and how they are tuned;
COUNTERPOINT_AND_VOICE_LEADING.md section 7 owns interlock as counterpoint;
REPETITION_AND_VARIATION_FORM.md section 5 owns cyclic variation form;
HOOK_AND_TENSION_ARCHITECTURE.md owns what makes a head cell memorable. The siblings decide the
ARCHITECTURE of a cue; this document decides its SURFACE — what happens inside one statement of a
cycle, at the scale of a single note and its neighbours — and in almost every tradition in our slice
the surface is where the intelligence lives.
than our canonical era. Usable, with the date stated in the sentence.
rather than dressed as evidence. Ch 2's deep horizon, Ch 10 and the two fairy-realm nodes are where
this tier does its work.
Ornament is usually taught as a list of squiggles. For a symbolic composer that is the wrong cut. The
cut that matters is what the ornament DEMANDS OF THE RENDERER, and there are four kinds.
| class | code | what it is | what it needs from the renderer |
|---|---|---|---|
| Note-level insertion | N | extra discrete notes added around a structural note — grace, mordent, turn, flick, melisma | nothing new; notes go into the note list |
| Formula substitution | F | a whole span of the line is replaced by a stored, named figure selected by its destination pitch | nothing new; a lookup table and a note list |
| Pitch-continuous | P | the pitch moves without ever resting — oscillation, slide, deflection, tension-drum glide | a per-note pitch envelope at high time resolution, which the round-2 path does not have |
| Timbral / mechanism | T | the pitch content is unchanged and the PRODUCTION changes — laryngeal register switch, damping, breath continuity, mute, rattle | either a second sample source or CC-controlled articulation, both partially available |
Two consequences set this document's recommendation order.
complexity in the slice. Javanese cengkok, Balinese kotekan figuration, Ethiopian melekket formulas
and the West African lead drummer's indexed phrase library are one computational object: a table of
named figures, each keyed to a goal tone or a metric position, substituted for a span of the
skeleton. Round 2 played a six-note head straight because nothing in the composer knew that in most
of these traditions a written note is an INDEX INTO A LIBRARY rather than a sound to be produced.
fairy-realm nodes — the slice's headline cue and its most sophisticated regions. Section 17 has the
work order.
harness/music_gen/pass2_realise.py:508-517 is the entire MIDI writer on the round-2 path: it appends
pretty_midi.Note objects with pitch, start, end and velocity and nothing else — no pitch_bends, no
control_changes, no channel management. Every round-2 cue is a bare note list, so class N and class
F are the only classes the generated corpus has ever been able to express, and it has used only the
plainest part of class N. The capability is not absent from the repository:
harness/music_gen/complete_track.py:1853-1872 carries _bend_events, interpolating a bend envelope
on a one-eighth-beat grid against an SFZ default range the code's own comment states as plus or minus
two semitones, and write_score_midi writes CC11 expression per part. That code is on a different arm,
has never fed a slice cue, and on inspection is not adequate as it stands (section 17).
Every worked example below transforms the SAME six notes, so the reader hears the difference the
tradition makes rather than the difference the tune makes. The six notes are round 2's own head cell,
verbatim from harness/music_gen/pass2_flores.py:87-88: MIDI 62, 66, 69, 78, 71, 69 — D4, F sharp 4,
A4, F sharp 5, B4, A4 — durations one, one, two, two, one, one beats across two bars of four-four, over
SCALE = [0, 2, 4, 7, 9] on D, which is D E F sharp A B.
Three properties matter. It is anhemitonic, so it maps onto the pentatonic frames of the Aka, San,
Ethiopian tizita and Tamil Mohanam systems without alteration. It contains one wide leap, a minor ninth
from D4 to F sharp 5, which several of these traditions treat as a marked event rather than as an
interval. And it is played straight, which is the defect Josh graded.
The skeleton is a TEST OBJECT, not a proposal — nothing here suggests fourteen regions share one tune.
It is held constant only so that the transformation is legible.
T0_Chapter_Index line 4 sets the era as a composite of Homo floresiensis at roughly 60,000 to
100,000 BP and pre-colonial Flores at roughly 1000 to 1500 CE. Two strata, two completely different
answers, and one of them is a void.
At the deep horizon there is no ornament vocabulary because there is no attested music. DECLARED
INVENTION is the only honest tier, and the instruction is to build the surface from acoustic mechanism
rather than from any culture's grammar. The one device defensible without inventing a tradition is
INTONATION DRIFT — a held tone that does not stay put, because holding a pitch steady is a trained
skill and the absence of training is a sound, not a silence.
At the pre-colonial stratum the melodic surface belongs to the gong cycle rather than to a tune. The
region page states the theory it has, "the gong-waning interlocking-cycle structure; the ceremonial
drumming that scores the caci rounds" (flores_island.md:855-856), and its own limit: the caci detail
and the gong-waning musical theory "are thin in the processed corpus" (flores_island.md:871-874).
External search confirms the ensemble shape and not the grammar ([Komodo Guide, Manggarai
Tribe](https://www.komodoguide.org/local-communities/manggarai-tribe/)) and returns no ornament
vocabulary at all. The entry is CLASS-ATTESTED ensemble, ABSENT ornament theory.
Deep horizon: the six notes are not played. One voice on the first pitch, entering under the ambient
floor, drifting between roughly forty and eighty cents around its centre over eight seconds, breath
audible at both ends, no second pitch until the cave's reverberation has answered the first. The
melodic content is one note; the interest is the instability and the room. Class T, renderable today
only as a sample choice, because we cannot bend.
Pre-colonial: the skeleton is re-read as a cycle marker. D and A become the two cycle-closing gong
strokes and the four notes between them become the gendang's business. Nothing is ornamented, because
the region has no attested ornament vocabulary and inventing one here would be a pastiche of a
neighbouring island's practice.
Period honesty. Any ornament grammar at the deep horizon is invention, and any ornament grammar at the
pre-colonial stratum borrowed from Bali or Java is a different culture's.
Care, quoted: "Represent-with-care posture: the Manggarai, Ngada, Nage, Lio, and Keo are protected
communities; combat never routes through them" and "Care means weight and authenticity, never
omission" (flores_island.md:489-495); and on music, "No reproduction of protected sacred music — the
gong-waning register is reference-composed, never sampled from consecrated performance"
(flores_island.md:856-857). The reframe that keeps the material: the gendang is held sacred in
Manggarai society, so the ensemble is composed at reference register and the caci rounds are scored as
a contest with real stakes rather than as ceremonial ambience — the weight stays, the consecrated
performance is not reproduced.
Pipeline. Classes: T only. Renderable today: yes. This node needs no engine capability and it needs an
acquisition.
T0_Chapter_Index line 5 sets a composite from pre-colonial Manggarai through the Portuguese 1500s
and Dutch colonial period.
The ornament answer is the same absence as Ch 2's later stratum. What canon ratifies is the caci duel
scored by the gong-waning at reference register, "rising at the caci duel" (flores_island.md:861-862),
so the melodic surface the record supports is RHYTHMIC-METRIC rather than melodic: the takitu
accompaniment rhythm, the graded gongs, the drum.
The transferable rule is design on top of the citation rather than inside it and is DECLARED as such:
the ornament of a caci cue derives from the duel's own action grammar, not from a melodic tradition. A
strike lands on the cycle; a parry displaces a stroke off it; the pair makes the figure. That is the
same operation Johnston documents for West African lead drumming as TIMING DISPLACEMENT (section 13.2),
reached here from choreography rather than from a drum tradition.
Worked transformation: the skeleton's six onsets are kept and its pitches discarded. Onsets at beats 0,
1, 2, 4, 6 and 7 become the gendang's phrase and the gong closes at beat 0 of the following cycle.
Ornament enters only as displacement — on the second statement the onset at beat 4 moves to 4.5 and the
onset at beat 6 to 5.75, so the figure is recognisably the same and metrically re-aimed. Classes N and
F in their rhythmic form; renderable today.
Period honesty. The colonial strata are inside this chapter's era range, so guitar string bands and the
modern brass-and-keyboard layer are period-plausible for the LATER windows and wrong for the earlier
ones. State the window per cue.
Care: the Ch 2 quotations bind identically (flores_island.md:489-495, 850-851). The
ornament-specific hazard is filling a documented gap with Balinese kotekan because Balinese kotekan is
what a composer knows — that would code Manggarai music as a lesser version of a neighbour's, which is
the flattening the care line forbids. The reframe is to score the gap as rhythm and leave the melody
sparse until the acquisition lands.
Pipeline. Classes: N, F (rhythmic). Renderable today: yes.
This is the slice's richest ornament node and the region page has already ruled that music here is
"not scoring but structure" (bali.md:1896).
subdivisions four or eight times faster than the core melody" (Wikipedia, Kotekan). The composite is
the line; neither performer plays it.
ornament for a sustained pokok tone is an oscillation between two adjacent scale degrees split across
two players.
note" of the figure; structurally the simplest, and difficult precisely because it is used at the
fastest tempos (Wikipedia, Kotekan).
one of its neighbor tones" (Wikipedia, Kotekan). This is the destination-first principle in its
purest documented form: the figure is generated by the pitch it is going TO.
"the middle pitch is always played in unison by both parts" (Wikipedia, Kotekan).
together (Wikipedia, Kotekan).
struck and then stopped on the knuckle, so the region's percussive signature is the SILENCE after
each note as much as the note. Get the damping wrong and the whole island sounds like the wrong
culture" (bali.md:2118-2120).
engineer the ombak beat-note" (bali.md:1923-1925), and the page requires it to emerge from two
sources with a controlled cents offset rather than from a chorus effect (bali.md:1937-1939).
The rules are strict and they are what make the practice intelligence rather than decoration.
A sustained pokok tone takes norot. A step takes nyok cok. Three available pitches take telu with its
obligatory shared middle; four take empat with no sharing at all. Choosing telu where the passage has
four pitches breaks the shared-middle rule and the composite stops interlocking.
it is the audible signature of the figure.
melody, the figuration has been written wrongly — the point is that the line exists only in the seam,
which the region page also treats as the chapter's thesis about art (bali.md:1957-1961).
ombak; and any figure applied to Rangda's material as a threat cue, because "Rangda is deliberately
never villainized" (bali.md:686) and "receives no monster sting, no low-brass motif, and no musical
resolution" (bali.md:2142-2144).
Before: D4 F sharp 4 A4 | F sharp 5 B4 A4, six notes, two bars.
Step one, reduce to pokok. The skeleton is far too fast to be a pokok. Keep four structural pitches at
one per bar of a four-bar cycle: D, A, F sharp (upper), A, with the last A on the gong stroke. The
six-note tune becomes a four-note skeleton — a REDUCTION, the opposite of what a Western
developing-variation approach would do first, and the correct first move here.
Step two, figure each pokok tone by its successor. D to A is a leap across the set, so the span takes
nyog cag with polos on beats and sangsih between them, each playing alternate notes of the D-E-F sharp-A
ascent. A to F sharp upper is a leap of a sixth, so the span takes nyok cok: both parts strike F sharp
in unison one subdivision early, then polos drops to E and sangsih rises to A and they alternate into
the arrival. F sharp back to A takes kotekan telu on F sharp, A and B, with A as the obligatory shared
middle struck by both. The final A is the sèlèh and takes norot only until the gong.
Step three, the timbral layer. Every stroke is damped, so the part is short events with composed
silence between them rather than a legato line. Each pair carries a fixed cents offset, a per-ensemble
parameter so a village ensemble and the Dalang's can differ (bali.md:1940-1942), and the shimmer
arrives as interference rather than as effect.
After: a four-note structural line carrying roughly sixty-four composite events across the same span,
produced by two players of whom neither plays the tune, wrapped in a beating no single source contains.
That is what "more intelligence in the melodies" means here, and none of it needs a pitch bend.
Period honesty. Gamelan gong kebyar, the idiom most listeners hear as Balinese, developed in north Bali
around 1915 — the organology lane names it as this node's headline anachronism trap
(ORGANOLOGY_SLICE_CULTURES.md:1583). Kotekan as a technique is far older than kebyar, but the
kebyar SPEED and its sudden-attack idiom are not the 1200 CE surface. Compose the figuration at the
slower ceremonial tempi.
Care, quoted: "the region renders at care tier elevated, which is the honest value for a living
religion with actively worshipped sites, living priesthoods, and named present-day custodian
communities — and elevated must not be allowed to trigger omission" (bali.md:679-682). On music
specifically: "the gamelan register is reference-composed, never sampled" (bali.md:2138-2139). The
ornament-specific care line, and it is a real one: the bajra and bajra-uter are consecrated implements
in use, so "their sound is recorded as design reference and reproduced by composition"
(bali.md:2145-2146) — meaning the rubbed friction tone may be composed and must not be a recording of
a priest's own instrument. The reframe that keeps the material intact rather than removing it: the
gender wayang's calming figure at the matatah is a real ritual function (bali.md:1912-1915), and it
is used in-game as the warding register at reduced density (bali.md:2059-2060) — witnessed, weighted,
and never turned into a combat cue.
Pipeline. Classes: F (the whole figuration system), N, T (damping, ombak). Renderable today: YES,
entirely, and this is the strongest argument in the document for building the formula-substitution
table before building the bend emitter.
The region page rules the routing before this lane opens: Thread 21 advances here at chapter-ambient
level, "the opposite of Ch 4, where music was the chapter's architecture" (sumatra_java.md:2231-2233),
and the substrate "carr[ies] almost no instrumental ethnography" while carrying "a fully named
vocabulary of the VOICE" (sumatra_java.md:2235-2236).
The ornament vocabulary this node actually earns is vocal and generic rather than melodic-atomic.
by women" (sumatra_java.md:2265-2266). The ornament here is LEXICAL. A Batak listener hears marked
language, so the genre's surface signature is a vocabulary, not a melisma, and a translation that
flattens it "loses the thing that makes it a genre" (sumatra_java.md:2275-2277).
(sumatra_java.md:2267-2268).
breaking", that is, circular breathing, with the sarune as the gondang sabangunan's main melody
bearer ([Aural Archipelago, Life, Death, and Gondang in North
Sumatra](https://www.auralarchipelago.com/auralarchipelago/gondang)). The taganing's tuned drum
melody "closely mirrors that played on the" sarune, which makes the region's melodic surface a
DOUBLED line across two families rather than a decorated single one.
The rule that follows, and it is unusual enough to be worth stating twice: this is the one node in the
slice where an unbroken melodic line is the authentic surface. The sarune does not phrase by breathing;
it phrases by figure. So the composed rest lives at the CYCLE level, between gondang pieces, and not
inside the line. That is a real tension with the rest-grammar axis that round one failed on, and the
resolution is that a Ch 5 sarune part is exempt from intra-phrase rest and the ensemble around it is
not.
Forbidden: the eighteenth-century Javanese court gamelan sound projected onto 1200, which the
organology lane names as this node's trap (ORGANOLOGY_SLICE_CULTURES.md:1584); and setting real tabas
or divination formulae as sung text, which the page rules out directly.
Before: the six-note skeleton with its three rests implied by the durations.
After, sarune reading: the six pitches are kept and every gap between them is filled, so the line runs
continuously across both bars. The one-beat D is approached from below and left by a filled step to F
sharp; the two-beat A becomes A with a passing E and back; the ninth leap to F sharp 5 is the only
event the line does NOT fill, because a leap that survives a filled context is the figure's one marked
moment. The taganing doubles the same contour on five tuned drums, one octave down in effect, so the
listener hears the same melody twice in two materials at once.
Before, six notes across eight beats with three silences. After, roughly twenty pitched events across
eight beats with no silence, doubled — and the leap is now audible as an event precisely because
everything around it has been filled.
Period honesty. Gondang sabangunan is Batak and its documentation is modern; the ensemble is
CLASS-ATTESTED for the culture and not dated to 1200. The page routes it as ambience for exactly this
reason. Say so on the cue rather than in a footnote.
Care, quoted: "the Batak, Javanese, Sumatran and Malay communities are protected and combat never
routes through them" (sumatra_java.md:703-704), and on music: "the pustaha's ritual CONTENT is
referenced, never reproduced... The voiced-inscription cue is composed to the sound of the register,
not lifted from a formula, and the gondang is reference-composed, never sampled from consecrated
performance" (sumatra_java.md:2570-2575). The ornament-specific hazard: the andung's special lexicon
is the property of a living genre performed at real funerals. The reframe is that the game composes an
andung-SHAPED lament with authored words in the register, cast as a named woman's role rather than
crowd ambience (sumatra_java.md:2277-2279), and reproduces no recorded lament.
Pipeline. Classes: N (the filling), T (breath continuity, and it is a real sample-selection
requirement — a sustaining double-reed patch with no re-articulation). Renderable today: yes.
This node's honest and distinctive finding is that its dominant register carries almost no melodic
ornament at all, and that its ornament intelligence lives in the drum syllabary instead. Pirit chanting
is described in the musicological literature as "a style of intoned recitation based on phonological
properties of the Pali language, but restricted melodically to the Ṛgvedic three-tone scale" (Lasanthi
Manaranjanie Kalinga Dona, "Healing through Pirit Chanting in Sri Lanka", Canadian Journal for
Undergraduate Research / cujucr, 2010 — the PDF at
https://www.cujucr.com/downloads/pdf_8_2010/Lasanthi.pdf returned HTTP 403 at time of writing and this
sentence is quoted from the indexed abstract, which is a weaker citation and is declared as such). A
three-tone recitation scale means one reciting tone with an upper and a lower inflection, with the
melodic decision made by the TEXT, not by the singer.
The drum side is where the vocabulary is. Kandyan practice teaches drum phrases through spoken
syllables — the bol or bera pada — and structures phrases through call-and-response between lead
drummer and ensemble, with the geta bera as the up-country core (summarised in [Music of Sri
Lanka](https://grokipedia.com/page/Music_of_Sri_Lanka) and [Insight Guides, Kandyan
Drumming](https://www.insightguides.com/destinations/asia-pacific/sri-lanka/cultural-features/kandyan-drumming);
both are tertiary and the region page's own substrate flag applies). The region page states that flag
plainly: "the Tovil rite's own drumming, its instrument names, and the Sinhalese chant tradition's
musical structure are absent from the processed corpus" (sri_lanka.md:951-954).
The rule: in a chant layer, pitch change is a function of syllable position, so ornament that arrives
for musical reasons is wrong by construction. In a drum layer, the ornament is a spoken syllable made
audible, so the figure must be pronounceable — a phrase a drummer could not say is not a phrase.
Before: six notes.
After, chant layer: three pitches. D becomes the reciting tone and carries most syllables; C sharp
below and E above are the two inflections, used at phrase-medial and phrase-final positions
respectively. The skeleton's melodic profile disappears entirely and what survives is its RHYTHM,
because the syllable count sets the rhythm and nothing sets the tune. The two-beat A becomes two
syllables on the reciting tone; the ninth leap becomes a single upward inflection at a clause boundary
and nothing more. The result is deliberately, correctly less melodic than the input, and it is the
strongest anti-exotica move available in the slice.
After, drum layer: the same six onsets become a spoken phrase, and the ornament is the ornamentation of
speech — a doubled stroke where the syllable doubles, a damped stroke where the syllable closes. The
lead drummer states it and the ensemble answers a reduced version, which supplies the two-layer stem
split for free.
The one melodic cue this node earns is canon's own: "a four-voice convergence for the peak ascent in
which each tradition's station adds one line and none of the four dominates the mix"
(sri_lanka.md:944-946). Four lines, four traditions, no ornament hierarchy — the ornament decision
there is to give each line the same density so none reads as the exotic one.
Period honesty. The horanava's shawm lineage read back into the Anuradhapura era, and sitar and tabla
anywhere in this chapter, are the traps the organology lane names (ORGANOLOGY_SLICE_CULTURES.md:1585).
The Kandyan drumming tradition as documented is up-country and later than Anuradhapura; treat it as
CLASS-ATTESTED and say the window on the cue.
Care, quoted: "the Sinhalese, Vedda and Tamil communities and the Buddhist sangha are protected; combat
never routes through them" and "Care means weight and authenticity, never omission: the wars stay in at
full weight" (sri_lanka.md:483-484, 497-498). On music: "Reference-composed, never sampled from
consecrated performance, and the protective chant is carried at reference register with no reproduction
of liturgy" (sri_lanka.md:925-926). The ornament-specific hazard is the Vedda material, because the
community "has been exhibited and written about more than it has been listened to"
(sri_lanka.md:488-490) — so the Vedda register is canon's own "breath and single struck notes rather
than melody, never scored as exotica" (sri_lanka.md:938-939), and adding ornament to it to make it
more musically interesting would be precisely the defect the page names.
Pipeline. Classes: N (minimal), F (drum syllable phrases). Renderable today: yes.
This is the node where the ornament IS the music, and where our renderer cannot currently produce the
music at all.
Gamaka is the general term for the movement applied to a svara. The classical treatises differ on the
count — Sarangadeva describes fifteen, Narada nineteen, Haripala seven ([Wikipedia,
Gamaka](https://en.wikipedia.org/wiki/Gamaka_(music))) — and modern practice works with a smaller
functional set, of which the following are the ones a composer must implement. The precise physical
definitions below follow the Subbarama Dikshitar and Gordon Swift taxonomy as reported in the search
literature; the primary Swift PDF at carnaticcorner.com returned HTTP 403 at time of writing and this
is declared.
motion and rarely rests at either extreme. This is the ornament that defines the largest number of
ragas.
rather than struck.
upper neighbour and pressed down onto it.
above. The repeated-versus-non-repeated distinction is what separates this pair from nokku and
odukkal and it is not decorative — it is what makes a repeated note not sound repeated.
The load-bearing fact, and the one that answers Josh's grade most directly: a raga is not a scale. The
Wikipedia article states it as "the unique character of each raga is given by its gamakas", and names
ragas that cannot exist without a specific one — Nayaki, Sahana, Devagandhari and Yadukulakambhoji
require kampita (Wikipedia, Gamaka). Two ragas can share an identical pitch set and be different ragas
because different degrees are oscillated. A pitch-set-only composer literally cannot write this music.
The placement rule is precise and it is a gift to a symbolic composer, because it is mechanical.
"Sa and pa, as consonant tonic–fifth, never feature gamaka deflections" ([Michael Schachter,
"Structural Levels in South Indian Music", Music Theory Online
21.4](https://www.mtosmt.org/issues/mto.15.21.4/mto.15.21.4.schachter.html)).
the direction of deflection follows a "magnetic pull" toward the structural tones: flattened degrees
deflect downward toward their lower neighbour, natural or raised degrees deflect upward (Schachter,
MTO 21.4).
type persists across tempi and changes its extent (Schachter, MTO 21.4).
changes the raga; and importing a Hindustani meend of a fourth or fifth into a Carnatic context,
because Carnatic deflection is characteristically narrow and anchored while Hindustani meend is wide
and connective. Hindustani terms are named here for contrast only — meend, the smooth glide between
notes; andolan, the gentle oscillation reserved for microtonal degrees; kan, the subtle grace note
that links main notes and without which, as one teaching source puts it, the melody stops sounding
Indian; murki, a fast delicate turn; khatka, a crushed grace cluster leading into a note ([Raag
Hindustani, Ornamentation](https://raag-hindustani.com/Embellishment.html)) — and none of them
belongs in a Tamil chapter set before the northern classical tradition took its present shape.
Map the skeleton to a raga first. D E F sharp A B on D is Sa Ri2 Ga3 Pa Dha2, which is Mohanam, the
Carnatic pentatonic that our anhemitonic set already is. This is not a coincidence to hide; it is the
reason the transformation is legible.
Before, as svaras: Sa Ga Pa | Ga(upper) Dha Pa.
After:
else. It is the only fully stable event in the phrase.
roughly plus or minus 80 to 150 cents at about five oscillations per second, never resting at either
extreme. On a fretless string or in the voice this is one continuous gesture; written as notes it is
a lie.
the oscillating Ga — so the ARRIVAL is continuous even though the destination is still.
nokku: F sharp 5 is attacked from A above and pressed down onto it, so the ear hears the upper note
bent into place rather than struck.
Count the classes. Two of six notes are stable pitches. Four carry continuous pitch motion. The
resulting pitch trace across two bars has two plateaux and four curves — which means that as a note
list, this phrase does not exist. It exists only as a pitch envelope, and that is the exact shape of
the capability gap in section 17.
Period honesty. The violin entered Carnatic practice around 1800 and the nagaswaram in a Sangam scene
is the other trap the organology lane names (ORGANOLOGY_SLICE_CULTURES.md:1586). Sharper for this
lane: the gamaka VOCABULARY as a named, enumerated taxonomy is a treatise tradition, and the treatises
that give us the names span roughly the thirteenth century (Sarngadeva) to 1904 (Subbarama Dikshitar).
Sangam-era Tamil practice had pann, the modal system tied to the poetic landscapes, and the yazh; the
continuity between pann and later raga is real and the modern gamaka taxonomy is not attested for the
Sangam window. The honest formulation for a cue: the PRACTICE of deflecting unstable degrees toward
stable ones is old and the NAMES are later, so use the practice and do not put the names in an NPC's
mouth. Ancient Tamil sources state the pann-to-raga continuity in terms of "shared swara progressions
and gamaka ornamentations" (Ancient Tamil music),
which supports the practice claim and not the taxonomy claim.
Care, quoted: "The Tamil and Malayali communities, the kalari lineages, the Siddha tradition, the temple
and its practitioners, the smiths and healers, and the honored Raghavendra are protected non-targets,
and combat never routes through them" (south_india.md:406-408). On music: "the page reproduces no
formula, no chant text and no consecrated performance. The command cadence is composed at reference
register from its documented FUNCTION... Protected sacred music is referenced, never sampled"
(south_india.md:800-803). The ornament-specific case: Vedic chanting has its own accent system and
reproducing it would be reproducing liturgy. The reframe that keeps it: the chapter's chant presence is
carried as REGISTER — the acoustic fact of a large room of trained male voices on a narrow pitch band —
with no Vedic text set and no accent pattern transcribed, while the gamaka vocabulary above is applied
to the chapter's SECULAR melodic material, where it is a craft and not a rite.
Pipeline. Classes: P dominant, N secondary, F absent. Renderable today: NO. This node is the primary
justification for the section 17 work order.
The region page opens Section 10 with the constraint: "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" (swahili_coast.md:840-843). This lane's contribution is therefore
mostly a dating discipline, and the dating is stark.
Taarab — the melismatic, modally Arab-inflected Swahili vocal music that a listener recognises
instantly as "Swahili coast" — originates in the late nineteenth century under Sultan Seyyid Barghash
bin Said, who reigned 1870 to 1888, sent the musician Mohammed Ibrahim to Egypt to study classical Arab
music, and on his return founded Zanzibar's first taarab orchestra; the oldest surviving ensemble,
Ikhwani Safaa, dates from 1902 ([AramcoWorld, The Soul of Zanzibar's Taarab
Music](https://www.aramcoworld.com/articles/2025/the-heart-moving-sound-of-zanzibar); [African Music
Library, Taarab](https://www.africanmusiclibrary.org/genre/Taarab)). That is roughly five hundred years
AFTER our canonical era. A taarab surface in a 1200-1400 CE Kilwa scene is the single largest
period error available in this region, and it is the error a well-intentioned composer will make,
because it is what the research returns first.
What IS period-plausible, and the register is different:
and the choice of maqam varies with time of day, place, teacher and mosque ([Middle East Eye, The art
of the adhan](https://www.middleeasteye.net/discover/adhan-muslim-call-prayer-melodies-maqams)). The
ornament rule here is genuinely precise and genuinely transferable: melisma is placed where the
TEXT'S phonology prolongs, under the tajwid rules of madd, and not where a musician feels a
cadence. Ornament governed by text rather than by taste is exactly the intelligence Josh is asking
for, and it is the one Islamic-practice rule this document extracts.
qasidas written by Hadrami scholars sit inside coastal Islamic performance to this day, with maulidi
performance built of qasidas "sung in a medley-like fashion, one merging into the next" ([Brill,
Maulidi Celebrations on the Northern Swahili
Coast](https://brill.com/edcollchap-oa/book/9789004522626/BP000021.xml?language=en)). The
medley-of-praise-poems FORM is the citable structure; the specific modern repertoire is not
fourteenth century.
sampled from a consecrated performance" (swahili_coast.md:858-860).
Honest note on the maqam layer, because it is easy to overclaim: maqamat are melodic modes rooted in
classical Arabic and Persian music that were "developed and refined as part of musical traditions,
later named and categorized" (Middle East Eye, above). The named-and-categorised system postdates our
window. So the transferable content is flexible intonation and text-governed melisma, and NOT a named
maqam on a cue sheet.
Before: six notes, evenly ornamented, meaning not ornamented.
After: the six notes become the syllable-bearing skeleton of a sung Swahili line, and melisma is
allotted by prolongation rather than by beat. Suppose the line's third syllable carries the vowel that
prolongs. Then D and F sharp take one note each; A — the prolonged syllable — takes five to seven
notes over its two beats, descending and returning, and never leaving the pentatonic set; the ninth
leap to F sharp 5 takes one note because it falls on a short syllable and a leap on a short syllable is
a marked event; B takes two; and the final A takes three with a slight downward drift at the end,
because the phrase closes. The line is now unevenly ornamented, and the unevenness is determined by
language.
Class N throughout, and it is fully renderable today. Melisma is the cheapest large melodic-complexity
win in the slice after Balinese figuration.
Care, quoted: "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" (swahili_coast.md:417-420),
and "Care means weight and authenticity, never omission" (swahili_coast.md:424).
The ornament with a genuine care hazard is the adhan, and the hazard is specific: the call to prayer is
a live act of worship performed by a named person five times a day, and generating one — synthesised,
sampled or composed — puts a fabricated act of worship in the player's ears. The reframe that keeps it
rather than removing it, and canon already carries the shape: the page treats the call as "cutting
across" the market bed as "a diegetic oneshot that the score gives way to" (swahili_coast.md:864-866).
Keep it diegetic, keep it distant, and record it as a real human performance under the voice-tiering
gate rather than generating one. The grove at the kaya edge is handled the same way — near silence,
"the score stepping back for real forest, which is also the care posture" (swahili_coast.md:868-869).
Pipeline. Classes: N dominant; P for flexible intonation, which is unrenderable and which the honest
alternative is simply to omit rather than approximate with a quantised bend. Renderable today: mostly.
This node has the slice's best documented musical theory and, uniquely, it has TWO ornament systems
that must never be mixed.
Ethiopian Orthodox chant runs in three modes attributed to St Yared: gəʿəz, the simplest, for ordinary
days; ʿəzəl, gentle, for fasts and funerals; and araray, the most ornamented, for great festivals
(Wikipedia, Ethiopian chant; [Sewasew, Saint
Yared](https://en.sewasew.com/p/saint-yared-(%E1%89%85%E1%8B%B1%E1%88%B5-%E1%8B%AB%E1%88%AC%E1%8B%B5))).
Mode selection is by OCCASION, not by mood — which is the same rule the region page already found in
its own source: performance "varies with the occasion — from simple recitation to elaborate
cantillation according to the character of the feast" (ethiopia.md:745-747).
The notation is the ornament system. Melekket comprises roughly 558 signs; each sign sits over a
syllable of text and cues a specific stored melodic formula, and "in performance, a formula is
embellished with improvised melodic ornaments" (summarised from Britannica's melekket and Ethiopian
chant entries via search indexing; the Britannica page returned HTTP 403 at time of writing and this is
declared). The malakkat are described as the minimal structural unit, learned in training and strung
together into longer melodies.
That is class F in its purest documented form, and it is structurally identical to Javanese cengkok:
a written mark is an INDEX into a memorised library of figures, and the performer's art is selection
plus a further layer of improvised surface. Two unrelated traditions in our slice, six thousand
kilometres apart, both solve melodic richness the same way — which is a strong argument that our
composer should implement the mechanism once and re-voice it.
The secular modal system is qenet, whose principal modes are tizita, bati, ambassel and anchihoye
(with yematbela wex sometimes counted fifth). They are "rendered with flexible intonation and
ornamented, melismatic singing", generally untempered on traditional instruments, and the azmari
performance surface is characterised by "vocal ornaments, slides, and agile turns; interweaving
instrumental fills between vocal lines", typically over a drone on the tonic ([Melodigging,
Tizita](https://www.melodigging.com/genre/tizita); [Melodigging,
Azmari](https://www.melodigging.com/genre/azmari); [Chromatone, Ethiopian
Kinits](https://chromatone.center/theory/scales/ethiopian/)).
The rule that separates the two systems and must not be broken: zema is liturgical, notated,
mode-by-occasion and performed by the debtera, whom the region page identifies as simultaneously
"the musician, the scholar, the scribe, and the warder-off of demons" (ethiopia.md:752-754). Qenet is
secular, unnotated, ornamented by improvisation and performed by the azmari. Applying azmari slides to
a liturgical cue, or setting a qenet song in melekket formulas, is not a stylistic slip — it crosses a
real institutional line inside a living church.
Tizita's pitch content is an anhemitonic pentatonic, so our skeleton maps onto it directly with no
alteration: D E F sharp A B is tizita-shaped on D. Once again the fit is the point, not a coincidence
to hide — the round-2 head cell was already in an Ethiopian-compatible set and sounded nothing like
Ethiopia, which proves the whole thesis of this document. The pitch set was never the problem.
After, secular qenet reading:
measurably off the tempered grid, and on a krar or masenqo that is simply where the string is.
a fixed reference and the intervals become the melodic content.
taken clean, so it registers as the phrase's event.
leaves the note and returns — and the turns are not identical, because in an improvised surface
identical ornaments read as a machine.
practice is built on, which means the cue is never a single line even when it is a solo song.
After, sacred zema reading: the same six pitches become six syllable positions. Each takes a formula
of three to eight notes drawn from a fixed library, selected by the mode of the day, and strung. The
six-note phrase becomes roughly thirty notes, and — the crucial part — a DIFFERENT thirty notes on a
festival than on an ordinary day, because araray and gəʿəz draw different formulas for the same signs.
That is adaptive music described in a sixth-century tradition, and the region page already noticed it:
"the same service sounds different on different days by rule" (ethiopia.md:747-748).
Period honesty, and this node has the sharpest instance in the slice. The CHANT is attributed to Yared
in the sixth century; the WRITTEN NOTATION is later, with one scholarly view placing the hymnary's
signs under the sixteenth-century emperor Claudius; and the tradition's own careful position is that
what Yared evolved was "not the chant itself but its graphic expression" (ethiopia.md:738-742). The
region page keeps the three positions separate and this lane keeps them separate too. Consequence for
the build: an Aksumite scriptorium scene may not show melekket notation on a page. It may show chant.
The scriptorium boss is a Zagwe-window figure and the notation question routes to the window the cue
is set in.
Care, quoted: "The Ethiopian Orthodox Tewahedo faith, its clergy, its debtera, its faithful community,
the Beta Israel, and the Agaw and Amhara highlanders are protected communities; combat never routes
through any of them. The faith is luminous throughout" (ethiopia.md:380-383), and on music: "the
region's zema is drawn from the documented tradition and no protected sacred music is sampled,
reproduced, or imitated in a consecrated form, and no liturgical text is set" (ethiopia.md:764-766).
The ornament-specific care line, and it needs a real reframe. A melekket formula library IS liturgical
music. Populating one from transcriptions of the Deggwa and rendering it would be reproducing the
chant, which the care line forbids in its own words. The reframe that keeps the intelligence without
crossing the line: implement the MECHANISM — a sign-indexed formula library with occasion-selected
variants and an improvised surface layer — and populate it with authored figures composed in the
mode's interval habits, never with transcribed Yaredawi formulas. The player experiences a service
that sounds different on a feast day because the system is real; no verse of the Deggwa is sung. That
is the care doctrine's own instrument: weight and authenticity, achieved by building the structure
rather than by copying the content.
Pipeline. Classes: F dominant (both systems), P (flexible intonation, untempered degrees, slides), T
(the sistrum's rattle and the prayer-staff clap as separate layers,
ethiopia.md:726-729). Renderable today: F yes, T yes, P no. Note that untempered degrees are not
strictly a bend problem — a per-instrument Scala tuning file solves fixed microtonal offsets without
any CC — and that is a cheap, separable win worth taking before the bend work.
rift_valley.md:626-632 states the position before anything else can be said: "Documented instruments:
none attested in this corpus, and the section says so before it says anything else... Because no
musicological source is in the corpus, no rhythmic structure, scale, call pattern or instrument is
described anywhere on this page."
This lane adds nothing to that and must not. What it can do is state the ornament consequence
precisely, and separate the two strata the chapter actually contains.
vocabulary for Olduvai at 1.8 million years because there is no music, no instrument and no attested
voice. Anything a composer writes there is invention, and the correct register is canon's own: "a
deep-time silence at Olduvai — near-absence with a floor of wind and stone" (rift_valley.md:637-638).
The melodic surface is the absence of one, and the ornament class is T — the sound of a place, not of
a performer.
dance clapping and the women's song-line rising at the n/um dance, carried from ratified canon at
reference register" (rift_valley.md:627-629). Two ornament facts follow, and both are structural
rather than melodic. The clapping is the ornament layer — dense polyrhythmic handclap is a
first-class part in this repertoire, not a garnish. And the "rising" in the song-line is the
documented gradual upward drift discussed at Ch 13 below, which is a whole-piece pitch behaviour and
not a note-level device.
Period honesty, stated plainly as the brief requires: a bone flute at Olduvai would be roughly 1.76
million years early (ORGANOLOGY_SLICE_CULTURES.md:1589). The earliest uncontested musical instruments
are tens of thousands of years old, not millions. Every deep-time sound in this chapter is
RECONSTRUCTION of an acoustic environment and DECLARED INVENTION of any human sound in it, and the cue
record should carry both words.
Care, quoted: "The Hadza, the Sandawe, the Maasai, the /Xam, the //Gwi and the Ju|'hoansi are protected
communities... Care means weight and authenticity rather than omission — and on this page the honest
form of weight is refusal to invent, because the material that would carry the weight is not here yet"
(rift_valley.md:367-372). On music: "No protected sacred music is reproduced, sampled, transcribed or
notated, and the reference register means a composed cue that honours the shape of the practice, never
a recording or a reconstruction of it" (rift_valley.md:629-631). The healing dance is a live rite; the
reframe that keeps its weight is the page's own — the panel cue is "built as the dance's counterweight
rather than as horror: pressure without menace-scoring" (rift_valley.md:645-646).
Pipeline. Classes: T only. Renderable today: yes. No ornament work is authorised for this node until
the acquisition lands, and that is a finding rather than a deferral.
COUNTERPOINT_AND_VOICE_LEADING.md section 7.1 and REPETITION_AND_VARIATION_FORM.md section 5.4
already carry the Aka four-part architecture and its variation grammar, and this lane does not repeat
them. What it adds is the SURFACE fact those documents leave open: in this repertoire the primary
ornament is not a pitch decision at all.
constant alternation between laryngeal mechanisms" — what singers call chest and head voice — with
"melodies with mainly wide intervals and using specific vowels correlated to two yodel registers"
(Susanne Fürniss, "Diyei and yeli. Yodeling in two musical cultures of Central Africa", Analytical
Approaches to World Music Monograph Series vol. 1, 2021, CC-BY,
Zenodo, pp. 68-98).
"take a shortcut"; and kuka ngo dikukè, a rhythmic-splitting process (Susanne Fürniss, "Aka
Polyphony: Music, Theory, Back and Forth", HAL —
the full text was access-blocked at time of writing and this is quoted from indexed abstracts, a
weaker citation, declared as such). That is decisive for our classification: the tradition itself
files a TIMBRAL operation in the same category as its rhythmic and substitutional ones. Our
four-class table is not an imposition here; it matches the culture's own taxonomy.
change" (Zenodo record, above) — a half-yodel, which is an ornament INTENSITY control and exactly the
kind of graded parameter an adaptive score needs.
"overlapping entries, a held drone under a descending cascade, a hard collective cutoff, and a
listened-for reply from the environment" (central_africa_congo.md:1029-1035).
skeleton already complies, which is why it can be used here at all.
is where the register break naturally falls, and this is a placement rule, not a preference.
yodel that keeps one vowel across the break is audibly wrong.
staged as a horror cue, because the theology has no evil in it"
(central_africa_congo.md:1077-1078).
Before: six notes, one voice, one timbre.
After: the skeleton is assigned to a dìyèí-type part and the leap becomes the ornament. D4, F sharp 4
and A4 are sung in mechanism one — chest — on a back vowel; the ninth leap to F sharp 5 crosses the
break and the upper note is sung in mechanism two on a front vowel, so the arrival is a change of
INSTRUMENT and not merely of pitch; B4 stays in mechanism two; and the final A4 drops back through the
break, which is audible as a second, softer event. The pitch content is unchanged. The line is now
unmistakably this repertoire and not any other.
Second layer, from the women's chorus brief: three voices enter one after another on the same figure,
each holding her entry note while the next comes in, so the descending cascade sits over an accumulating
drone; all stop together on one beat; and a composed silence of about two seconds follows for the
forest's echo, with the ambient bed briefly unmasked. The hard cutoff and the silence are the region's
musical signature and they cost nothing to implement.
Period honesty. The likembe is a nineteenth-century instrument spread by colonial labour migration
(ORGANOLOGY_SLICE_CULTURES.md:1590) and has no place in a pre-colonial forest cue. The Fang secular
dance sequence the region page documents — belebele, elong, ball with its accordions and guitars,
manzang — is explicitly a MODERN-STRATA lane (central_africa_congo.md:1056-1060) and its instruments
must not leak backward.
Care, quoted: "Every community named in this section is protected; combat never routes through any of
them... Care here is elevated and elevated means depth: the hard material is carried, not smoothed"
(central_africa_congo.md:552-555).
Two ornament-specific care items. First, the Bwiti ngombi harp is a body — the Sister of God's backbone,
ribs, sinews and stomach, into which her voice comes during the ceremony (central_africa_congo.md:1040-1045).
The reframe: the harp's ornament vocabulary is not ours to author, and the region page's own line holds
— "referenced and never reproduced, and never scored as intoxication" (central_africa_congo.md:1082-1083).
Second, and it is a hard gate rather than a caution: "the Baka music lane rests on canon assertion
without region-shelf substrate and must not be composed until the acquisition lands"
(central_africa_congo.md:1093-1095). This section is therefore a REQUIREMENTS document for that lane
and not a licence to compose it.
Pipeline. Classes: T dominant (the mechanism switch), N secondary. Renderable today: PARTIALLY, and the
honest description of the partial is important. A mechanism switch can be approximated by swapping to a
second sampled voice mid-line, and our renderer already emits one stem per part, so the swap is
expressible. It is a two-performer fake of a one-performer act, and the cue record should say so.
The region page calls this "one of the page's richest lanes rather than one of its thinnest"
(west_africa.md:2400-2401), and the ornament answer here is the most conceptually distant from
Western practice in the whole slice.
The dùndún is a tension drum whose pitch is controlled by squeezing the cords. Yoruba has three level
tones plus glides between them, and the drum tracks them: dùndún drumming "allows for manipulation of
gliding pitch contours in ways that correspond to the differentiation of the Yorùbá linguistic tone
levels", and the acoustic study finds that the drum "very accurately mimics the microstructural
characteristics of Yorùbá vocalization" while fidelity decreases when the drums are used for music
rather than direct speech (Durojaye et al., "When Music Speaks: An Acoustic Study of the Speech
Surrogacy of the Nigerian Dùndún Talking Drum", [Frontiers in Communication,
2021](https://www.frontiersin.org/journals/communication/articles/10.3389/fcomm.2021.652690/full)).
Four modes are distinguished — rhythm, singing, drum talking-performative, and drum talking-direct —
which gives us a four-position dial between pure music and pure speech.
The ornament is therefore the GLIDE, and it is not decorative: it is a phoneme. A dùndún line rendered
as discrete pitches is not a simplified dùndún line; it is a line with its consonants removed.
The region page supplies the ensemble rule that surrounds it. At Oyo dawn music, "all secondary
instruments normally used for dance are excluded except the gudugudu, a small kettle drum, which alone
supplies the basic pattern against which no fewer than five talking drums play poetry in unison"
(west_africa.md:2422-2425). One invariant pattern, five voices in unison on a poetic text: that is a
complete orchestration specification and it is not a texture any Western scoring habit would produce.
The igbin family gives an explicit ornament instruction that is worth quoting because it is a
tension-architecture device: in a three-drum set the two smaller drums "give intermittent rhythmic
phrases dropped almost as soon as they become clearly established in the listener's mind, leaving him
suspended in the middle of the phrase" (west_africa.md:2450-2453). Deliberate incompleteness, as
practice.
The structural rule of the family: the bell holds and the lead varies. In Ewe practice the support
drums have some room for variation "except for the bell, which is completely fixed" ([Wikipedia, Ewe
drumming](https://en.wikipedia.org/wiki/Ewe_drumming)). The lead drum has "the largest set of prescribed
phrases and greatest leeway for variation".
The lead drummer's operations are documented and named, and they are directly implementable. Jeryl
Johnston, "Variation Techniques in Four West African Lead Drumming Examples: A Comparative Study",
AAWM Journal 11/2 (2023), covering
Gahu (southern Ewe), Adowa (Akan), Agbadza (Ewe) and dùndún (Yoruba), identifies four:
figure always appearing AT THE END. Destination-first, again, and stated by an independent tradition.
fundamental. This is the density dial, and note that emptying is on the list: taking notes away is a
variation technique, not an absence of one.
independent repetition and variation.
Forbidden: varying the bell; giving the Ogboni drums a voice; scoring any of the protected communities
with an antagonist cue.
Before: six pitches.
After, and the first move is the radical one: the pitches are discarded. The dùndún does not play a
scale; it plays an utterance. The skeleton's six events become six syllables of an oriki line, and the
melodic contour is derived from that line's tone pattern — high, mid, low and the glides between them.
Where our skeleton leapt a ninth, the utterance has a low-to-high transition, and the drum executes it
as a CONTINUOUS SQUEEZE across roughly 120 milliseconds rather than as two notes. Where our skeleton
repeated A, the utterance has two mid tones, and the second is distinguished from the first by attack
rather than by pitch.
Then the ensemble: the gudugudu holds one invariant pattern; five dùndún play the oriki contour in
unison; the two smaller igbin drums enter with phrases and drop them before they resolve. Then the four
Johnston operations across successive cycles — the phrase is amplified over two periods with the
original figure landing at the end of the second; the following cycle displaces the middle motive by
one pulse; the next fills the space between the second and third onsets; the next segments the tail
into a two-note cell and repeats it three times.
Six notes have become a spoken sentence, played by five drums over an unvarying bell, varied by four
named operations across four cycles, in which the melodic content is a continuous pitch trace and not a
note list at all.
Period honesty. The djembe is Mandé and its global spread is mid-twentieth century
(ORGANOLOGY_SLICE_CULTURES.md:1591). The gbedu-type pedestal drums "almost certainly predate the
dundun" (west_africa.md:2453-2455), so a strictly Ile-Ife-classical scene has a different drum roster
from a Dahomey-1700s scene, and both are inside this chapter's era band — state the window per cue. The
rada tradition and the ring shout are Ch 42 and Ch 43 material and the page has already held them here
(west_africa.md:2478-2486); this lane holds them too.
Care, quoted: "Ch 12 renders at ELEVATED care (CVD §6.2)... The Yoruba, Fon, Dogon and Akan communities
are protected communities... Care means weight and authenticity and never omission, so nothing above
softens what follows" (west_africa.md:1155-1161). On music: "no protected sacred music is reproduced.
The Ogboni drum carving only cult members may see, the Fon ancestral songs whose texts the
ethnographer himself could not obtain, the Dogon sigu so, and the Jos Plateau instruments the community
classes as sacred are all referenced at form and register — never sampled, never recreated, never given
an authored ceremonial text" (west_africa.md:2569-2573).
The ornament with the sharpest care problem is oriki itself, because oriki is not generic praise: it is
genealogy, jurisprudence and social claim at once (west_africa.md:2431-2437), aimed at real lineages.
The reframe that keeps it — and it is the design's best moment rather than a compromise — is that the
oriki the player hears is aimed at the PLAYER, is authored for the game's own fictional lineage, and
changes with what the player has done. No real family's oriki is drummed. Herskovits could not obtain
the Fon ceremonial texts and recorded that he could not (west_africa.md:2472-2476); the build inherits
his limit and composes the FORM — circular movement toward the drums, overlapping chorus entry, hard
collective refrain change, lead-and-chorus response (west_africa.md:2483-2485) — with no ceremonial
text authored.
Pipeline. Classes: P dominant and unrenderable — the glide is the whole ornament; N for the filling
operation; F for the indexed phrase library. Renderable today: NO for the headline device. Note the
special difficulty: a dùndún glide is a per-note pitch envelope on a PERCUSSION source, which our
sample-based renderer cannot bend at all without a bendable tension-drum instrument that the palette
does not contain and that the care line would in any case gate.
The region page's finding governs the ornament answer: the archive contains a song CATALOGUE whose
"organizing principle is that a song belongs to a specific creature, person or state"
(south_africa.md:601-602), naming the Cat's Song, the Song of the Caama Fox, the Songs of the Blue
Crane, the Old Woman's Song, the Song of the Bustard, the Song of the Springbok Mothers, ||kabbo's Song
on the Loss of his Tobacco Pouch, the Broken String, and others (south_africa.md:603-606).
That makes the melodic surface REFERENTIAL before it is decorative. A short melody's job here is to
name a thing. So the primary ornament rule is a restraint rule, and the region page has already written
it: "sparse, vocal-first, rattle and struck-idiophone percussion, no orchestral pad, and a deliberate
absence of harmony under the melody" (south_africa.md:613-614).
The documented vocal-surface devices, all CLASS-ATTESTED from Ju|'hoansi and broader San recording
rather than from the /Xam archive specifically, and the distinction matters:
shifts — the same timbral-ornament class as Ch 11, reached independently ([Cultural Survival, The
"Success" of San
Music](https://www.culturalsurvival.org/publications/cultural-survival-quarterly/success-san-music);
[NTS, British Library Sound Archive: Ju|'hoansi healing songs from the
Kalahari](https://www.nts.live/editorial/british-library-sound-archive-healing-songs)).
scale, stretching and pushing higher throughout a song" (Cultural Survival, above). This is the
slice's one documented WHOLE-PIECE pitch ornament, and it is unusual enough to be a signature.
|goin|goin named in the archive itself (south_africa.md:606-608).
The Great Zimbabwe leg brings a different and later system, and it must be kept separate. Shona mbira
practice pairs kushaura, the leading part in the higher range, with kutsinhira, the interlocking
response in the lower, over a cyclical pattern containing multiple intertwined melodies and extensive
room for melodic and rhythmic variation ([Wikipedia, Shona
music](https://en.wikipedia.org/wiki/Shona_music); [MBIRA, Shona Mbira
Music](https://mbira.org/what-is-mbira/mbira-music/shona-mbira-music/)).
Before: six notes.
After, /Xam catalogue reading: the six notes become a NAMED short song of about four seconds, attached
to one creature, and its ornament budget is almost nil — one clap layer at twice the song's density,
one rattle, and no harmony. The one ornament applied is the documented upward drift: over eight
repetitions the tonal centre rises by roughly a semitone in total, continuously rather than in steps,
so the eighth statement is audibly higher and no single repetition sounds like a modulation. That is
the whole transformation, and it is deliberately the least ornamented answer in this document, because
this is the region where more ornament would be the error.
After, Great Zimbabwe leg: the six notes become a kushaura cycle in the upper range; kutsinhira enters
in the lower range displaced by half a pulse, playing the same cycle's complementary onsets; the
composite reveals two or three melodies that neither part contains, which is the tradition's own
described effect and the same emergent-line principle as Balinese kotekan reached independently.
Period honesty, and it is sharp here. The modern mbira dzavadzimu presented as the fourteenth-century
instrument is the organology lane's named trap for this node (ORGANOLOGY_SLICE_CULTURES.md:1592) and
the ornament version of that trap is worse: kushaura and kutsinhira as a NAMED paired practice is
documented in the modern ethnographic record, principally through Paul Berliner's 1970s work, and is
CLASS-ATTESTED rather than dated to the dry-stone florescence. Use the interlocking principle; do not
claim the fourteenth century for the terminology. The /Xam archive itself was recorded in the
nineteenth century from a specific set of speakers, which the page states in its care line, so even the
catalogue is a nineteenth-century record of an older tradition and not a window onto it.
Care, quoted: "Every community here is protected; combat never routes through any of them... Care means
weight and not omission, so the hard material stays: the archive was recorded from men who were
prisoners" (south_africa.md:390-393). On music: "no sacred ceremonial content is reproduced. The
trance-dance and rain-rite music is referenced and never recorded, performed or synthesized, and no
melody attributed to a living ceremonial tradition enters the score" (south_africa.md:620-622).
The ornament-specific care item is the healing-dance repertoire, and the reframe is precise: the
documented upward drift and the interlocking-clap texture are STRUCTURAL properties, and structural
properties are not melodies. Using the drift and the interlock in an authored, non-ceremonial song is
building with the tradition's grammar; transcribing a healing song is not, and the page forbids it in
its own words. The named-song catalogue solves the same problem elegantly on the discovery side — "a
player who learns a creature learns its song" (south_africa.md:611-612) — with authored songs in the
archive's FORM.
Pipeline. Classes: T (register shifts, clap density), P (the upward drift). Renderable today: PARTIALLY.
A semitone-quantised drift is trivially renderable by transposing successive cycles; the CONTINUOUS
drift the source describes is not, and a quantised approximation would read as a key change, which is
the wrong effect. This is a second, independent argument for the section 17 capability.
T0_Chapter_Index lines 2 and 3 give region_primary as Fairy Realm and era as "Pre-game timeframe
(dimensional)". There is no ethnography, and the organology lane has already ruled the palette and the
prohibition: build the register from an acoustic MECHANISM rather than from a culture, and do not
borrow the slice's own devices — "The gong cycle, the kotekan interlock, the hocket, the talking drum
and the drone are Earth's, each belonging to a specific living people"
(ORGANOLOGY_SLICE_CULTURES.md:1500-1504).
This lane's job is to say what the MELODIC SURFACE is when no tradition may supply it, and canon already
contains the answer in a place the music lane has not yet read. The Prologue's SQ7 seed is
"the resonant-pitch sealing, the arc's first resonance craft, and its repetition until the pitch-finding
is a habit rather than a feat" (docs/spine/CH_PROLOGUE.md:195).
That is an ornament specification. The fairy realm's melodic surface is THE PITCH SEARCH: a tone that
begins away from its target and glides continuously onto it until the room answers, with the arrival
signalled by resonance rather than by a cadence. It is diegetic — the player is learning to find a
pitch, and the score is doing the thing the player is learning — and it belongs to no living people,
because it is a physical act rather than a cultural one. DECLARED INVENTION, and grounded in canon's
own quest seed rather than in a composer's whim.
Three further constructive rules, applied rather than asked:
Bali damps, West Africa strikes, the Congo breaks register. The realm slides and never re-attacks.
That single contrast makes the two worlds separable by ear before any instrument is identified, and it
costs nothing.
come from two real sources with a controlled offset (bali.md:1937-1939). The realm therefore may not
carry a tuned two-source ombak. What it may carry, because it is a different phenomenon, is
INHARMONIC PARTIAL CLASH from a stretched octave — non-periodic, spectral, arising from the tuning
rather than from a pair — which the organology lane already specifies as the realm's tuning approach.
This is an execution call made inside the locked vision and it is named here so it can be vetoed.
foreshadow cue anywhere, least of all before CHPRO_B14" (CH_PROLOGUE.md:245) and "NO prophetic or
foreshadow cue anywhere in the ending sequence — no dread swell under the chase, no held note on the
cord coming off, no minor turn before the rift (a cue that warns the player is the chapter's single
most damaging failure)" (CH_01.md:238). For this lane that means the pitch-search ornament, which is
by nature a device of anticipation, must always RESOLVE in the origin material. An unresolved glide is
a warning, and a warning here is the named failure.
Worked transformation. Before: six notes. After: the first note is not struck. It enters roughly a
minor third below D, on wine glass and bowed vibraphone, and rises continuously over about three
seconds until the room's sympathetic resonance blooms on D — the arrival is a TIMBRAL event, an increase
in sustained partial energy, not an attack. F sharp is found the same way but faster, because the seeker
is learning. A is found instantly, because by the third attempt the pitch-finding is habit, which is what
the quest seed says. The ninth leap is the one place the search fails and overshoots, arriving a little
above F sharp 5 and settling down onto it. B and the closing A are struck cleanly, because the skill has
landed. The melody has become a competence curve, and the player hears themselves getting better before
they are told they are learning anything.
Care. There is no living tradition to protect here and therefore no care line to quote from a region
page — which is itself the finding. The binding constraints are canon's own: "no reproduced
living-tradition liturgy" (CH_PROLOGUE.md:245, CH_01.md:238), and the theme registry's realm deny
rule that a realm register may never be built by exoticising a living human tradition
(ORGANOLOGY_SLICE_CULTURES.md:1481-1482). The care operation for these two nodes is therefore
subtractive in one direction only: keep every Earth grammar out.
Pipeline. Classes: P dominant, T secondary. Renderable today: NO. The arc's FIRST CUE — the player's
first thirty seconds of the game — is built on the one ornament class the renderer cannot produce. That
is the strongest single argument in this document for the section 17 work order, and it should be
weighed as such.
Five ornament devices appear in four or more of the fourteen nodes, which means one implementation
serves many palettes.
(Wikipedia, Cengkok); Balinese nyok cok generated from the
NEXT pokok pitch (Wikipedia, Kotekan); Ethiopian melekket signs cueing stored formulas strung toward
a cadence; the West African lead drummer's amplification, in which the original figure "always
appears at the end" (Johnston, AAWM 11/2); the Sri Lankan drum syllabary's pronounceable phrases.
Five traditions, one mechanism.
the raga; the Yoruba dùndún glide, where a discrete pitch removes a phoneme; Ethiopian qenet's slid
arrivals; the San upward drift; the fairy realm's pitch search. Five nodes where a discrete-pitch
rendering is not a simplification but an error.
shifts; the Fon ancestral voices as a falsetto-and-groan pair (west_africa.md:2513-2515); Balinese
damping, where the silence after the stroke identifies the culture. Four nodes in which who is playing
is signalled by HOW rather than by WHAT.
Carnatic condensation and expansion across tempi; Balinese figuration density by irama level. Four
nodes that raise intensity by ornamenting more, never by adding a pad — which is also exactly what
layer_mode: swap already means in the machine cue tables.
placed by madd prolongation; Sri Lankan pirit pitch chosen by syllable position; Batak andung's marked
lexicon; Balinese speech levels the dalang commands. Six of fourteen nodes treat melody as language,
which is a strong argument for deriving contour from text rather than from a scale wherever a text
exists.
ORNAMENT IS APPLIED BACKWARDS FROM ITS DESTINATION, NEVER FORWARDS FROM ITS ORIGIN. Fix the goal tone
and its metric position first; choose the figure that lands on it; then decorate the notes inside that
figure — and leave the goal tone itself plain.
Every tradition in this slice states some version of it independently. Balinese nyok cok generates its
figure from the pitch it is going to. Javanese cengkok are selected by sèlèh. Carnatic gamaka deflects
toward the structural tones and Sa and Pa never take gamaka, so the destination is exactly the note
that stays clean. West African amplification develops over a multiple of the period with the original
figure at the end. Ethiopian formulas are strung toward a cadence. Aka parts are positioned relative
to a cycle boundary that none of them articulates together.
Two corollaries a composer can act on immediately:
Sa, the cadence and the collective cutoff are all plain. A cue that decorates its arrivals is
ornamenting in the wrong direction, and that is the specific defect that makes pastiche.
surface. Round 2 wrote a tune and then looked for somewhere to put it. Every tradition here does the
reverse.
The claim in the lane brief — that our renderer emits no CC curves — is confirmed for the round-2 path
and needs one honest amendment.
harness/music_gen/pass2_realise.py:508-517 writes notes only: pitch, start, end,velocity. No pitch bends, no control changes, no channel management. Every class P ornament in this
document is currently unrenderable on the path that produced the graded cues.
harness/music_gen/complete_track.py:1853-1872 interpolates a composed bend envelope and
write_score_midi writes CC11 expression per part. It has never fed a slice cue, and on inspection it
would be insufficient if it did.
Why the existing emitter is insufficient, stated precisely so the fix is scoped correctly:
_bend_events interpolates on a one-eighth-beat grid. At a typical cue tempo thatis on the order of 70 to 80 milliseconds per point. A Carnatic kampita oscillates at roughly four to
seven cycles per second, so an eighth-beat grid supplies about two points per oscillation — a
triangle wave, not a gamaka. A per-note pitch envelope sampled at roughly 200 Hz is the requirement.
so a monophonic ornamented line is fine; any layer carrying simultaneous notes cannot give them
independent bends without one channel per voice or an MPE-aware path.
(complete_track.py:1856). That is adequate for gamaka, which is characteristically narrow, and
inadequate for a Hindustani-width meend or a dùndún glide across the full tone span. Wider bends
require bend_up and bend_down opcodes on the instrument, and a repository grep for those opcodes
returns nothing, so no instrument in the palette currently declares a non-default range.
bends its transient too, which is audibly wrong. This one is not fixed by CC at all; it needs either a
bendable tension-drum instrument, which the palette does not contain and the care line would gate, or
a synthesis path.
The work order, in the order it should be taken, because the order is itself the finding:
figures, each with an interval shape and a destination degree, plus a selector that picks a figure
from the current structural pitch and the next one. That single feature delivers Balinese figuration,
Javanese cengkok, Ethiopian melekket, the West African indexed phrase library and the Sri Lankan drum
syllabary. It is pure score-side, it is checkable by a predicate over the notes in the same way the
existing one-deviant-per-head rule is, and it is the largest melodic-complexity gain available in the
slice.
non-equidistant slendro and pelog, and the fairy realm's stretched octave are all fixed pitch offsets,
not moving ones, and a Scala file per instrument solves them with no CC path at all.
bend_envelope field in the plan, an emitter that writes bend events at roughly 200 Hz, one MIDI channel per ornamented voice, and bend_up/bend_down declared on
the instruments that need more than two semitones. This unlocks Ch 7, the Ethiopian secular surface,
the San drift and both fairy-realm nodes.
problem and not a CC problem.
Canon, and it outranks everything below it: docs/spine/DECISIONS_PENDING_JOSH.md (ROUND 2 GRADED,
line 5600 following); registries/T0_Chapter_Index [ACTIVE v1.2]; docs/spine/CH_PROLOGUE.md;
docs/spine/CH_01.md; and Section 10 plus the represent-with-care paragraph of flores_island.md,
bali.md, sumatra_java.md, sri_lanka.md, south_india.md, swahili_coast.md, ethiopia.md,
rift_valley.md, central_africa_congo.md, west_africa.md, south_africa.md. Sibling proposal-tier
lane: docs/proposals/music/craft_research/ORGANOLOGY_SLICE_CULTURES.md. Code read rather than assumed:
harness/music_gen/pass2_flores.py, harness/music_gen/pass2_realise.py,
harness/music_gen/complete_track.py, harness/music_gen/sfz_palette.py.
Scholarly and reference, with their tier marked:
https://www.mtosmt.org/issues/mto.15.21.4/mto.15.21.4.schachter.html — peer-reviewed; the source for
the Sa-and-Pa-never-deflect rule and the magnetic-pull account.
Study", AAWM Journal 11/2 (2023),
https://journal.iftawm.org/previous/2023-volume-11-no-2/johnston/ — peer-reviewed; the four named
lead-drum variation techniques.
Speech Surrogacy of the Nigerian Dùndún Talking Drum", Frontiers in Communication (2021),
https://www.frontiersin.org/journals/communication/articles/10.3389/fcomm.2021.652690/full —
peer-reviewed; the glide-as-tone-level finding and the four modes.
Series vol. 1 (2021), pp. 68-98, CC-BY, https://zenodo.org/records/5608397 — peer-reviewed; the
laryngeal-mechanism definition and the half-yodel strategy. The record page was reachable; the full
PDF was not, and the technique detail is quoted from the record's own description.
— the four named parts and the Aka's own variation vocabulary. The full text was ACCESS-BLOCKED at
time of writing and these claims come from indexed abstracts and figure captions, which is a weaker
citation and is declared as such.
https://www.cujucr.com/downloads/pdf_8_2010/Lasanthi.pdf — the three-tone recitation-scale claim.
The PDF returned HTTP 403 and the claim is quoted from the indexed abstract; declared as weaker.
Shona music, Ewe drumming. Tertiary; used for terminology and figure definitions, corroborated
against at least one further source wherever a rule is stated.
https://www.auralarchipelago.com/auralarchipelago/gondang — field documentation by a working
ethnomusicologist; the masiulak hosa circular-breathing detail.
Barghash dating, consistent across both.
art of the adhan" — the qasida-medley form and the maqam-by-occasion practice.
healing songs from the Kalahari" — the upward-drift and interlocking-voice descriptions. Both are
popular secondary sources; the drift claim deserves a specialist source before it hardens, and
Nicholas England's book-length Ju|'hoansi study is the named acquisition target.
contrast set, which is explicitly excluded from the slice.
the melismatic-with-drone performance description.
Offered to the pages that own them; NOT canon edits.
south_india.md Section 16's music acquisition sharpens to a gamaka performance-practice sourcerather than Tamil temple music generally — the ornament taxonomy is what a composer cannot proceed
without.
swahili_coast.md Section 16's music acquisition should be re-scoped: the period-correct vocal layeris Islamic recitation and Hadrami qasida practice, not taarab, and a taarab acquisition would buy a
nineteenth-century answer to a fourteenth-century question.
ethiopia.md:774-777 records the musical-theory substrate as GOOD. This lane adds that a melekketformula library populated from transcription would cross the page's own sacred-music line, so the
mechanism is buildable and the content must be authored.
flores_island.md:871-874 and central_africa_congo.md:1093-1095 are confirmed from the ornamentside: neither node has an ornament vocabulary in the corpus, and Ch 11's is gated from composition.
T0_Theme_Registry has no field expressing ornament class. If the section 17 formula library is built, a per-cue ornament_class_list over the four codes would let a gate check that a cue's
declared ornament is one the renderer can actually produce.