THE KONDO SCHOOL — the Nintendo dialect, rule by rule
CANON SUBORDINATION — this document is PROPOSAL-TIER: it serves canon and never outranks it.
The canon: the CVD · the T1 foundation docs · the T0 registries (registries/) · the spine
(docs/spine/CH_*.md) · the region pages (_source/02_Tier_2_Region_Pages/). Authority order: docs/DOC_MAP.md § 0.
Canon served: NONE FOUND — this document cites no canon anchor anywhere in its body.
Do not build from it alone: open the spine node (docs/spine/CH_NN.md), the region page
(_source/02_Tier_2_Region_Pages/), and the T0 registry your work targets, and derive from those.
READ THAT CANON FIRST — open it and derive from it before you build anything from this document.
If this document disagrees with canon, CANON WINS and this document is the defect — fix the
document, never the canon. Nothing here is applied until it is ratified into canon.
The grammar of the exemplar class Josh named — Super Mario Bros., Super Mario 64, A Link to the Past, Ocarina of Time, Majora's Mask — stated as rules a composition can be checked against rather than as appreciation. It is the school that supplies the anti-fatigue law (never close the loop with a perfect authentic cadence), the four-step phrase model, and the four-mutually-unlike-motifs mechanism Kondo states in his own words.
---
0. Provenance — where this text came from, and what was done to it
TIER: PROPOSAL. This is research, not canon. It binds nothing until ratified; the CVD and the
T1/T0 canon outrank every line of it.
SOURCE: the research return of workflow run wf_d1ba57a8-e47, agent a4077c9299e20ffe6, 2026-08-05. The return was a
structured object with three keys — rules, evidence, measurable_features — and all three are
reproduced below.
EXTRACTION: VERBATIM and mechanical. The strings were read out of the run journal by script and
written here unedited — not re-worded, not re-ordered, not summarised. Only the headings, the
bullet markers and this provenance section are new. That matters because NOSTALGIA_RUBRIC.md
cites these rules BY ID, and a citation that resolves to a paraphrase of its rule is the same
WHY IT EXISTS AS A FILE: the returns lived only in the run journal, which is agent scratch and
is not canon-readable. The rubric quoted them anyway and said the mechanism was "stated verbatim
from the school grammars" — which named documents that did not exist. That is finding F3 of the
cold verification of 2026-08-05, and this file is half of its fix (the other half is the rubric's own
provenance line).
WHAT IS NOT CLAIMED: that every source below was independently re-verified. The evidence list
carries the citations the research agent gave, with its own declared gaps and negative controls
left in. Verifier finding F4 stands open against this: several named scholarly sources are bare
author-year strings, and a reader cannot resolve them from this repo alone. Treat a citation here
as a pointer to check, never as a checked fact.
---
1. The rules
MELODY-01 OPENING INTERVAL INSIDE A PERFECT FIFTH: the head interval of the hook must fall within P5, and should cluster on M2/m3/M3. Across every Kondo cue catalogued in Ferguson's JMTP corpus the opening interval is <= P5, with none at a 6th or 7th: OoT Title = M2 up; SMB Underwater = M2 up; LoZ Overworld / Song of Healing / Astral Observatory = M2 down; Song of Storms / Rosalina in the Observatory = m3 up; SM64 Main Theme / SMB Invincible = m3 down; Saria's Song = M3 up; SMB Overworld / NSMB Overworld / SMW Athletic / Beedle's Shop = M3 down; SM64 Dire Dire Docks = P4 up; ALttP Hyrule Castle, ALttP Kakariko Village, OoT Shop = P4 down; SMB3 Overworld = P5 up. The wide-leap opening (6th, 7th, octave) is NOT in this dialect.
MELODY-02 THE THREE-NOTE HEAD CELL: build the theme's identity on a 3-pitch cell stated in the first two bars, in a long-short-long rhythmic shape, then reuse that exact cell as the seed of every later phrase. Zelda's Theme states its basic idea across mm. 1-2 as a three-pitch cell in a half-quarter-half rhythm; all six OoT ocarina songs are built on a repeated three-note motif. The cell, not the full phrase, is what a listener carries for 30 years.
MELODY-03 SMALL PITCH ALPHABET FOR SIGNATURE/DIEGETIC TUNES: cap the distinct pitch set of a player-facing signature tune at ~5 pitches. Kondo authored the OoT ocarina songs (Zelda's Lullaby, Saria's Song, Song of Storms et al.) against the five ocarina buttons: 'I had to create all of those memorable tunes with only five tones.' Constraint is the hummability engine, not an obstacle - the melodies survive whistling because they were designed for a five-hole instrument.
MELODY-04 ARCH CONTOUR, COMMON NOT EXOTIC: short phrases that rise are answered by a longer phrase that falls (Song of Storms, Saria's Song). Do NOT chase contour novelty. Jakubowski et al. (2017, Psychology of Aesthetics, Creativity, and the Arts) found involuntary-musical-imagery ('earworm') tunes have MORE COMMON global melodic contours than non-earworm tunes relative to a large pop corpus - the ordinary rise-then-fall shape is a positive predictor. Novelty belongs in the interval structure and the harmony, never in the contour.
MELODY-05 LEAP-THEN-FILL WITHIN THE EIGHT-BAR UNIT: contrast a leap-dominated half against a stepwise/neighbour half inside the same 8-bar statement. In 'Minuet of Forest' the first four bars use large octave leaps and the second four are 'much more compact, using only neighbour and stepwise motion until the rising arpeggiated cadence' (Gervais 2015). One gesture per half; never mix the two textures inside a half.
MELODY-06 ANCHOR ON SCALE-DEGREE 5: privilege scale-degree 5 as the melody's structural pole and returning tone. Named explicitly as a 'Kondo-ism' in Everitt's Queen's University style study (Inquiry@Queen's 2024), alongside parallel motion, chromatic harmony, modes, Picardy third, and the bVI-bVII-I cadence.
MELODY-07 OUTLINE THE TONIC TRIAD IN THE OPENING STATEMENT: the first melodic statement should arpeggiate or outline the tonic triad in a small range before any chromatic move. Gervais shows this across the OoT warp songs: 'Prelude of Light' opens outlining a D major tonic triad, 'Serenade of Water' outlines the D minor tonic triad in mm. 1-2 and repeats it in mm. 3-4, and Song of Storms is built on a minor arpeggio. His conclusion: each piece 'has a relatively small melodic range outlining a tonic triad.'
PHRASE-01 HOOK BY BAR 2, NO STANDING INTRO: the melodic identity must arrive within the first two bars. The Library of Congress National Recording Registry essay on the SMB Ground Theme (Gibson 2023) locates the identity in 'the first six notes that greet the player upon starting the game.' Ferguson's whole catalogue indexes Kondo cues by OPENING INTERVAL because the opening is the identity. A 4-bar atmospheric run-up before the tune is a defect in this dialect.
PHRASE-02 THE CAPLIN SENTENCE IS THE DEFAULT 8-BAR UNIT: 2-bar basic idea -> near-literal 2-bar repeat -> 4-bar continuation that fragments the cell -> cadence. Zelda's Theme runs two such sentences (mm. 1-8 and mm. 9-16, the second reusing the identical basic idea), then a third 8-bar sentence for the B section (mm. 17-24) whose basic idea is a descending fifth and whose continuation uses liquidation. SMB2 Overworld is catalogued as a Sentence in Ferguson's Form section. Gervais independently finds Caplin's 'repeating basic idea' in 'Prelude of Light.'
PHRASE-03 CALL AND RESPONSE ACROSS TWO TIMBRES: split the 8-bar unit into two 4-bar halves and give them different instruments, so the second half reads as an ANSWER. In 'Minuet of Forest' the first half imitates Sheik's lyre and the second imitates the player's ocarina - literally teacher-then-student (Gervais 2015). The timbre change, not a new melody, is what makes the repetition feel like dialogue.
PHRASE-04 REPEAT THE BASIC IDEA LITERALLY BEFORE DEVELOPING IT: mm. 3-4 must be a near-exact restatement of mm. 1-2. Zelda's Theme repeats its basic idea at mm. 3-4 and again at mm. 11-12; 'Serenade of Water' repeats mm. 1-2 in mm. 3-4. Development belongs in the continuation, never in the statement.
PHRASE-05 NEVER CLOSE THE LOOP WITH A PERFECT AUTHENTIC CADENCE - THIS IS THE ANTI-FATIGUE LAW: Gervais's four-step model of Kondo's phrase grammar ends with 'a weakened cadence that tends to end with a major tonic chord'; the formal analysis of Zelda's Theme independently identifies 'avoidance of the tonic note and perfect authentic cadences,' cadencing instead on half cadences, which is precisely what makes the looping structure keep pulling. A fully-closed cadence at the wrap point is the single most common way a 90-second loop becomes unbearable.
PHRASE-06 SIGNAL THE MOVE OFF TONIC PROLONGATION WITH RHYTHM + BASS DIRECTION, NOT VOLUME: at the pivot into the second half, change the melody's rhythmic pattern AND reverse the bass's direction of travel. Gervais on 'Minuet of Forest': 'The shift from tonic prolongation to predominant is signaled by the changes in the melody's rhythm and the bass motion, which moves downwards instead of upwards.'
HARMONY-01 THE FOUR-STEP KONDO PHRASE MODEL (the spine of the whole dialect): (1) a tonic prolongation, usually alternating I and IV over a held common bass; (2) a subdominant-function chord closing the prolongation; (3) a CHROMATICALLY ALTERED chord where the dominant is expected; (4) a weakened cadence landing on a MAJOR tonic. Derived by Gervais (2015) across 'Minuet of Forest,' 'Prelude of Light,' 'Serenade of Water,' and 'Requiem of Spirit.' Every element is negotiable per cue; the ORDER is not.
HARMONY-02 PROLONG TONIC BY ALTERNATING I-IV OVER A COMMON BASS FOR THE FIRST FOUR BARS: the held scale-degree-1 bass is the 'fundamental base' (Harrison) that makes I<->IV read as one tonic field rather than two chords. Explicit in 'Minuet of Forest' (I7 <-> IV, scale-degree 1 held) and 'Serenade of Water' (i <-> IV with D as the common bass tone).
HARMONY-03 SUBSTITUTE bVII FOR V AT THE CADENCE: bVII reads as dominant without a leading tone because it preserves the 4->3 motion into tonic. 'Serenade of Water' cadences IV - VI - VII - I instead of V-I (Gervais). This substitution is the mechanical cause of the weakened cadence in PHRASE-05.
HARMONY-04 THE bVI-bVII-I 'MARIO CADENCE' IS THE HOUSE TRIUMPH GESTURE: borrow both chords from the parallel minor and land on the major tonic. Named tracks: the SMB end-of-level fanfare, the SMB Ground Theme (it closes section B and forms the basis of section C), the SMB3 airship victory fanfare, SM64 'Dire Dire Docks' (I | bVII | I | bVII | bVI | bVII | I), and OoT ocarina songs. Everitt lists 'bVI - bVII - I' as a core Kondo-ism. Reserve it for arrival, resolution, and reward moments - it is the sound of a thing being achieved.
HARMONY-05 MODAL MIXTURE IS THE DEFAULT COLOR SOURCE, NOT AN ORNAMENT: borrow freely from the parallel mode in an otherwise major cue. Ferguson catalogues Kondo mixture chords by track: i and bIII in LoZ Overworld and SM64 Dire Dire Docks; iv in SMW Athletic, Dire Dire Docks, and SMB3 Ending; iio in SMB3 Overworld; a Neapolitan in LoZ Overworld. The Splice harmonic reading of the LoZ Overworld 8-bar loop names v6, bVI, bVII and a V/V inside eight bars in Bb - 'unfamiliar harmonic colors' at one chord per bar.
HARMONY-06 ONE DECLARED MODE PER CUE, AND SOUND ITS CHARACTERISTIC DEGREE: LYDIAN (raised 4, approached by step so it lands playful rather than dissonant) for wonder, forest, and otherworld - Saria's Song, the OoT Title theme, and Majora's Mask 'Song of Healing' are all catalogued as Lydian. MIXOLYDIAN / absent leading tone for pastoral-heroic - 'Minuet of Forest' has no raised 7 at the cadence and is best read as E Mixolydian. DORIAN for melancholy-without-darkness - 'Serenade of Water' uses natural 6 in minor (D Dorian). The mode is the cultural/emotional signature; declare it before writing a note.
HARMONY-07 PICARDY THIRD TO CLOSE MINOR AND MODAL CUES: end on the major tonic triad. Gervais identifies a tierce de Picardie closing both 'Serenade of Water' and 'Requiem of Spirit' (the latter resolving via a melodic appoggiatura into the parallel major); Everitt lists Picardy Third as a Kondo-ism. It is how a weakened cadence still feels like an arrival.
HARMONY-08 CHOOSE CHROMATIC CHORDS BY VOICE LEADING, NOT BY FUNCTION: the substitute for the dominant is whatever chord lets every voice move by one semitone into the tonic. In 'Prelude of Light' m. 6 carries a common-tone seventh chord rather than any dominant, and Gervais reads it via Tymoczko's parsimonious voice leading - 'each voice moves by a semitone' - which 'generates a sense of cadential arrival, despite the lack of a clear dominant chord.'
HARMONY-09 DESCENDING CHROMATIC BASS AS THE TENSION ENGINE: hold a pedal, then walk the bass down chromatically under a static or slow melody. Zelda's Lullaby sits on a C pedal for four bars prolonging dominant harmony, then the bass outlines a chromatic descent from C to A across mm. 1-7. Kondo's progressions are described as strong 'by virtue of voice leading - chromatic descending bass line - and melodic flow.'
HARMONY-10 HARMONIC RHYTHM AT ONE CHORD PER BAR IN THE THEME STATEMENT: the LoZ Overworld 8-bar loop has 'consistent harmonic rhythm; the composition changes chords exactly once every measure.' Fast harmonic rhythm buries the melody; slow harmonic rhythm makes a 60-90s loop feel static. One-per-bar is the exemplar-class default; halve it only for ambient/water/ethereal cues.
HARMONY-11 SECONDARY DOMINANTS STAY IN THE VOCABULARY: V/ii, V/IV, V/V, V/vi, viio/V, viio/ii, viio/iii, viio/VII, and even V/N are all catalogued in the Kondo set. SMW 'Athletic' alone carries V/ii, V/IV, V/vi and viio/V; SMB 'Underwater' carries V/N and a tritone substitution; Majora's Mask 'Astral Observatory' carries V/V and V/vi. Diatonic-only progressions are NOT this dialect.
HARMONY-12 EXTENDED TERTIAN COLOR AT THE SLOW END ONLY: add9, 11th and maj7#11 chords belong to the ambient/water/celestial cues, not to the running cues. Catalogued: SM64 'Dire Dire Docks' (add9, 11, maj7#11), Majora's Mask 'Astral Observatory' (maj7#11), 'Song of Healing' (maj7#11). This is the jazz-fusion inheritance Kondo names himself - Herbie Hancock in the LoC interview, and Sadao Watanabe plus T-SQUARE cited as influences on the SMB Ground Theme.
HARMONY-13 PARALLEL MOTION / PLANING AS A TEXTURE DEVICE: move whole chords in parallel rather than resolving each voice independently. Listed as one of the four named 'Kondo-isms' (parallel motion, chromatic harmony, modes, proximal voice leading) in Everitt's Queen's University style analysis of Ocarina of Time.
FORM-01 MULTI-MOTIF LOOPS ARE THE ANTI-FATIGUE MECHANISM - >=3 MOTIFS, TARGET 4, MUTUALLY UNLIKE: this is Kondo's own answer, verbatim, to the question of how he avoids sounding repetitive (Library of Congress interview, 12 April 2023): 'there are four motifs to the music and each of those motifs is unique. They don't sound like the other motifs, and that's why it doesn't sound too repetitious.' A loop that rides one motif for 90 seconds fails by construction, no matter how good the motif is.
FORM-02 ASYMMETRIC SECTION ORDER, NOT AABB: the SMB Ground Theme runs Intro - {ABC - ADCD}, with the intro material returning as the CADENCE of section C - a structure described as unusually complex for a Mario theme and as balancing symmetrical against asymmetrical sections deliberately. Plain strophic AABB reads as filler in this class.
FORM-03 CALIBRATED REPETITION DENSITY, NOT MAXIMUM REPETITION: Hunt's corpus study of 21,391 video-game pieces across 11 consoles (NES through GameCube, N64 included) found a strikingly stable mean repetition score of 62.79% with SD 3.49 - i.e. roughly 63% of one-bar clips are exact repeats of an earlier clip, and this figure barely moves across hardware generations. That is the target band. Well under it reads as through-composed and unmemorable; well over it reads as loop fatigue.
FORM-04 MODULAR PHRASE POOLS FOR LONG-DWELL SPACES: for open-field / long-traversal music, author a POOL of interchangeable phrases with compatible seams rather than a fixed loop. Kondo's own GDC 2007 example: the Ocarina of Time Hyrule Field daytime music selects from 12 phrases, stitched into an 8-phrase melody, so the piece stays coherent yet fresh across repeated exposure. Seam quality is the design problem (cf. Medina-Gray's modular-smoothness method: meter, timbre, pitch, volume, abruptness at every join).
FORM-05 THEMED LOOPS RUN AT LEAST 24-32 BARS: Zelda's Theme is a 24-bar binary (a 16-bar A of two sentences plus an 8-bar B). Below ~24 bars a themed cue cannot carry the 3-4 contrasting motifs FORM-01 demands. Small forms inside the loop are period and sentence; binary and ternary sit above them.
RHYTHM-01 SET DUPLE MELODY AND BASS AGAINST A TRIPLE-SUBDIVIDED PERCUSSION LAYER - Kondo names this as the source of forward motion in the most famous game theme ever written (LoC interview 2023): 'the melody and the bass parts are divided equally, but the rhythm hi-hat is divided into triple. And I think that the gap between the even distribution and the triplet distribution is what gives the music the feeling of moving forward.' The cross-rhythm is the propulsion; it is not a groove preset.
RHYTHM-02 A CLAVE-DERIVED PERCUSSION CELL UNDER A STRAIGHT MELODY: the SMB Ground Theme's percussion carries a Latin clave cell (two dotted quarters plus a quarter) against straight melodic notes, creating the tension that reads as bounce. The lineage is real and documented: the LoC essay places the 'tropical influenced, marimba-like melody' next to Yellow Magic Orchestra's exotica synthesis and Japanese city-pop's Latin/Caribbean roots, and Kondo himself cited Sadao Watanabe and T-SQUARE.
RHYTHM-03 DERIVE TEMPO AND RHYTHMIC CELL FROM THE PLAYER'S LOCOMOTION, NOT FROM THE SCENE'S MOOD: Kondo made the theme's rhythm match the physical feeling of running and jumping, playing the game repeatedly to fix the melody and tempo, and his GDC 2007 first principle is rhythm: 'If the music doesn't reflect the rhythm of the game, it might as well be piped in from a source outside the room.' His own SMB anecdote is the proof - his first, softer, more heartwarming melody did not fit the on-screen action and was fully rewritten as upbeat and 'action'-y.
RHYTHM-04 RESTS ARE STRUCTURAL, NOT GAPS: Kondo, on composing under memory limits - 'I aim to achieve the maximum expression with a minimal amount of sounds. I try to evoke something in the silence, in the absence of sound. Rest notes are very important to me.' A wall-to-wall melody line kills hummability; the ear needs the space to sing back into.
RHYTHM-05 TEMPO IS AN INTERACTIVE PARAMETER: build cues that can speed up, thin out, or swap layers on game state. Kondo's SMB hurry-up speeds the piece when time runs short; SMW mutes/enables percussion on Yoshi; SM64 swaps instruments by Mario's location (shore, water, cavern); NSMB moves enemies on chord stabs. Author the variation states with the cue, not after it.
VOICE-01 THREE TO FIVE SIMULTANEOUS PARTS FOR A SIGNATURE THEME - THIN IS THE STYLE, NOT THE LIMITATION: the SMB Ground Theme was written for two pulse channels, one triangle and one noise channel, with Kondo 'alternating melodies between the three melodic channels' rather than doubling them. Ocarina of Time's engine had far more headroom (4 sequence players, SEQ_NUM_CHANNELS 16 per player, 4 layers per channel, 64 preallocated sequence layers in the shipped audio code) and the writing still stays thin. Modern orchestral density is the fastest way to lose this dialect.
VOICE-02 ONE JOB PER VOICE: lead / bass-as-counterline / harmonic ostinato / percussion. In the SMB Ground Theme the triangle-wave bass functions as counterpoint, not as root-doubling, and the noise channel builds a full kit from kick, snare and hi-hat variants. If two parts are doing the same job, delete one.
VOICE-03 OPEN VOICING: spread the chord tones. Kondo, on why: the Famicom's square waves 'have a richer harmonic content than normal instruments,' so open voicing 'sounds much clearer.' The principle survives the hardware - harmonically rich timbres need spacing, not close-position stacks.
VOICE-04 BALANCE IS A COMPOSITION-TIME CONSTRAINT, NOT A MIX FIX: Kondo's GDC 2007 second principle - plan instrument selection, pitch register, volume and stereo position so that sounds are not 'gobbled up' by ambience, and check cohesion across the SET (he prepared 4-5 songs and presented them to directors as a group to show how they work together). Score the soundtrack as a palette, not track by track.
VOICE-05 ONE UNAMBIGUOUS SONIC IDENTITY PER PLACE: Kondo's two stated overarching philosophies, as recorded in the Library of Congress registry essay, are 'to convey an unambiguous sonic image of the game world' and 'to enhance the emotional and physical experience of the gamer.' A cue that could belong to two regions has failed the first test. Per-culture period instrumentation (the standing scoring law) is the vehicle for this rule, not an exception to it.
PROCESS-01 SCORE AGAINST THE EXEMPLAR CLASS, NEVER AGAINST PROXIES: the defect that produced the rejected 14 themes was measuring distinctness, motif recurrence and a hummability proxy with no comparator anchor. Every candidate must be feature-extracted and scored against a centroid built from the named exemplar set (SMB Ground Theme, SMB3 Overworld, SMW Athletic, SM64 Main Theme, SM64 Dire Dire Docks, LoZ Overworld, ALttP Hyrule Castle, ALttP Kakariko Village, OoT Title, Zelda's Lullaby, Saria's Song, Song of Storms, Great Fairy's Fountain, Hyrule Field, Gerudo Valley, Majora's Mask Astral Observatory, Song of Healing, Wind Waker Grandma's Theme), with per-feature distance reported, not a single scalar.
PROCESS-02 DISTINGUISH THE GRAMMAR FROM ONE ANALYST'S EXTENSIONS: Everitt's study separates the observed 'Kondo-isms' (parallel motion, chromatic harmony, modes, Picardy third, emphasis on scale-degree 5, bVI-bVII-I) from his own 'personal Kondo-isms' (percussive ostinatos, added dissonances, quartal stacking, book-end themes) which are HIS derived extensions, not Kondo's practice. Our rig binds on the first list; the second list is the model for how a composer legitimately extends the dialect into a personal voice - which is exactly what this project needs to do rather than pastiche.
---
2. The evidence
PRIMARY - Koji Kondo, interview with the Library of Congress, 12 April 2023 (National Recording Registry induction of the SMB theme). https://www.loc.gov/static/programs/national-recording-preservation-board/documents/koji-kondo-interview.pdf | Supplies verbatim: the four-unique-motifs anti-repetition mechanism; the duple-melody-and-bass vs triplet-hi-hat cross-rhythm as the source of forward motion; the standing mandate 'rhythms that are easy to get into and that have memorable melodies in order to make games more fun to play'; the 3-tone + 1-noise Famicom constraint; Herbie Hancock / jazz-fusion background.
PRIMARY (essay of record) - David Gibson, 'Super Mario Bros. theme -- Koji Kondo, composer (1985)', National Recording Registry essay, Library of Congress, 2023. https://www.loc.gov/static/programs/national-recording-preservation-board/documents/Super-Mario-Brothers-Theme_Gibson.pdf | Supplies: Kondo's two overarching compositional philosophies ('unambiguous sonic image of the game world' / 'enhance the emotional and physical experience of the gamer'); the 4-channel allocation (2 pulse + triangle + noise) with melody ALTERNATING between melodic channels; the tropical/marimba melodic character and the YMO / city-pop / Sadao Watanabe / T-SQUARE lineage; 'the first six notes that greet the player' as the identity.
PRIMARY - Koji Kondo, GDC 2007 keynote 'Painting an Interactive Musical Landscape', as reported by Game Developer and Nintendo World Report. https://www.gamedeveloper.com/game-platforms/gdc-koji-kondo-s-interactive-musical-landscapes and https://www.nintendoworldreport.com/feature/13118/koji-kondos-gdc-2007-presentation | Supplies: the three principles (rhythm, balance, interactivity); the Hyrule Field modular figure - 12 phrases randomly selected and stitched into an 8-phrase melody; the 4-5-songs-as-a-group cohesion practice; instrument/pitch/volume/position balance so sounds are not 'gobbled up'; SMW Yoshi percussion layering and SM64 location-based instrument swaps.
PRIMARY (translated) - Koji Kondo 2001 composer interview, shmuplations. https://shmuplations.com/kojikondo/ | Supplies: 'I aim to achieve the maximum expression with a minimal amount of sounds. I try to evoke something in the silence, in the absence of sound. Rest notes are very important to me.'; the open-voicing rationale for square waves; the 'music should fit the game' / interactivity-first stance.
PRIMARY (translated) - 'The History of Nintendo Game Music (1983-2001)', shmuplations. https://shmuplations.com/nintendogamemusic/ | Supplies: Kondo's SMB rewrite anecdote (the first softer, heartwarming melody did not fit the action and was replaced with an upbeat, action-y one); per-console channel budgets (Famicom 3 + noise, Super Famicom 8, N64 variable ~20 and dependent on graphics load); Totaka on phrase/pattern reuse discipline.
SCHOLARLY - Nicholas Gervais, 'Harmonic Language in The Legend of Zelda: Ocarina of Time', Nota Bene: Canadian Undergraduate Journal of Musicology 8/1 (2015): 26-42. https://ojs.lib.uwo.ca/index.php/notabene/article/download/6599/5323/12371 | The single most load-bearing source. Derives the four-step phrase model (tonic prolongation -> subdominant -> chromatically altered quasi-dominant -> weakened cadence on a major tonic) from close analysis of 'Minuet of Forest', 'Prelude of Light', 'Serenade of Water' and 'Requiem of Spirit'; identifies I<->IV prolongation over a Harrison fundamental base, VII-for-V substitution, the absent leading tone (E Mixolydian in Minuet of Forest), D Dorian in Serenade of Water, the common-tone seventh replacing the dominant in Prelude of Light via Tymoczko parsimonious voice leading, tierce de Picardie closes, lyre-then-ocarina call and response, Caplin repeated basic idea, and 'a relatively small melodic range outlining a tonic triad'.
SCHOLARLY (catalogue) - Brent Ferguson (with T.J. Laws-Nicola and Alan Elkins), 'Music Theory Examples in Video Game Music', Journal of Music Theory Pedagogy 34 (2020), art. 12. https://bura.brunel.ac.uk/bitstream/2438/27859/1/FullText.pdf | Per-track classification of the Kondo corpus by opening interval (the full <= P5 distribution quoted in MELODY-01), by mode mixture chord (i, iio, bIII, iv, Neapolitan by track), by secondary function (V/ii, V/IV, V/V, V/vi, V/N, viio/ii, viio/iii, viio/V, viio/VII by track), by modality (Saria's Song, OoT Title, Song of Healing = Lydian), by extended tertian harmony (add9, 11, maj7#11 by track), by tritone substitution (SMB Underwater), and by small form (SMB2 Overworld = Sentence).
SCHOLARLY - Dominic Everitt (supervisor Stephanie Lind), 'Analyzing Compositional Style in the Music of The Legend of Zelda: Ocarina of Time: A Study in Developing Musical Identity', Inquiry@Queen's Undergraduate Research Conference Proceedings, 2024. https://ojs.library.queensu.ca/index.php/inquiryatqueens/article/view/17415 | Names the 'Kondo-isms' explicitly: parallel motion, chromatic harmony, modes, Picardy third, emphasis on scale-degree 5, and bVI - bVII - I. Also separates the author's own derived extensions (percussive ostinatos, added dissonances, quartal stacking, book-end themes) - the model for extending rather than copying the dialect. Bibliography confirms the field's core texts (Summers 2016 and 2021, Schartmann 2015, Collins 2008, Lind 2016, Gervais 2015).
SCHOLARLY (corpus statistics) - Samuel J. Hunt, 'An Analysis of Repetition in Video Game Music', CSMC+MuMe 2020. https://boblsturm.github.io/aimusic2020/papers/CSMC__MuMe_2020_paper_10.pdf | 21,391 pieces across 11 consoles (NES/SNES/Megadrive/NeoGeo/PlayStation/Saturn/N64/PS2/Dreamcast/GameCube). Headline number for the rig: mean bar-level exact-repetition score 62.79%, SD 3.49, essentially flat across hardware generations - so ~63% is a style constant, not a hardware artefact. Also documents that song length and track count, not repetition, are what grew with storage.
SCHOLARLY (earworm science) - Kelly Jakubowski, Sebastian Finkel, Lauren Stewart, Daniel Mullensiefen, 'Dissecting an Earworm: Melodic Features and Song Popularity Predict Involuntary Musical Imagery', Psychology of Aesthetics, Creativity, and the Arts (2017). https://www.apa.org/pubs/journals/releases/aca-aca0000090.pdf | Earworm tunes show MORE COMMON global melodic contours (the rise-then-fall arch) and LESS common average gradients between melodic turning points, plus faster tempi, versus a large pop corpus. Grounds MELODY-04: ordinary contour + unusual interval detail, never the reverse.
SCHOLARLY (method for modular loops) - Elizabeth Medina-Gray, 'Analyzing Modular Smoothness in Video Game Music', Music Theory Online 25.3 (2019). https://www.mtosmt.org/issues/mto.19.25.3/mto.19.25.3.medina.gray.html | Provides the seam-scoring method (meter, timbre, pitch, volume, abruptness) needed to evaluate FORM-04 phrase-pool joins and loop wrap points; illustrated on The Wind Waker.
ANALYTIC - 'Zelda's Theme: Sentence Structure' and 'Themes and Leitmotifs: Zelda's Theme', High Score music-theory blog, March 2020. https://highscore.music.blog/2020/03/19/zeldas-theme-analysis/ | Full Caplin formal reading: binary |A B|, three sentences - basic idea mm. 1-2 (three-pitch cell, half-quarter-half), repeat mm. 3-4, continuation mm. 5-8 by fragmentation, half cadence on V7; second sentence mm. 9-16 reusing the identical basic idea; B-section sentence mm. 17-24 on a descending-fifth basic idea with liquidation. States the loop mechanism directly: 'avoidance of the tonic note and perfect authentic cadences.'
ANALYTIC - Jason Brame, '8-Bit Analysis: The "Mario" Cadence', gamemusictheory.blogspot.com, April 2011 (cited in the Queen's bibliography). http://gamemusictheory.blogspot.com/2011/04/mario-cadence.html | Names the bVI-bVII-I cadence and its occurrences: SMB overworld and end-of-level fanfare, SMB2, SMB3 airship victory fanfare, SM64 Dire Dire Docks (I | bVII | I | bVII | bVI | bVII | I), and OoT ocarina songs. Explains why bVII reads dominant - it retains the scale-degree-4-to-3 motion without a leading tone.
ANALYTIC - 'Inside The Score: The Super Mario Bros Theme', Video Game Music Shrine. https://videogamemusicshrine.com/inside-the-score-the-super-mario-bros-theme/ | Gives the Ground Theme's form as Intro - {ABC - ADCD} with the intro returning as the cadence of section C; C major/Ionian with the Mario cadence closing section B and founding section C; the clave-derived percussion cell (two dotted quarters plus a quarter) against a straight melody; the per-channel job assignment (pulse = melody+harmony, triangle = bass as counterpoint, noise = kit); the T-SQUARE 'Sister Marian' half-clave influence.
ANALYTIC - 'How the overworld theme of The Legend of Zelda takes us on a harmonic adventure', Splice blog. https://splice.com/blog/legend-of-zelda-overworld-harmony/ | Bar-by-bar harmony of the 8-bar LoZ Overworld loop in Bb: I, v6, bVI, III/I, bVII, vi/i, V/V, V - four borrowed/chromatic events inside eight bars, at exactly one chord change per measure.
ANALYTIC - 'A Lesson in Modes from The Legend of Zelda' and 'A Lesson in Harmony from The Legend of Zelda', ThinkSpace Education. https://thinkspace.ac.uk/blog/a-lesson-in-modes-from-the-legend-of-zelda/ | Saria's Song reaches the raised fourth through a passing note so the Lydian colour reads playful rather than dissonant; Song of Storms is arpeggio-founded with the sixth withheld from the melody and alternating major/minor in the harmony; and the contour rule - short upward phrases are answered by a long downward phrase (Song of Storms, Saria's Song).
SOURCE (Kondo quoted) - 'Zelda's Lullaby: A Five-Tone Melody', Marisa's Music Studio. https://www.marisamusicstudio.com/zeldas-lullaby-a-five-tone-melody/ | Quotes Kondo: 'I had to create all of those memorable tunes with only five tones of the classic do-re-mi scale' - the ocarina songs' pitch alphabet was fixed by the playable instrument's buttons. Grounds MELODY-03.
ENGINE (hard numbers) - zeldaret/oot decompilation, include/audio.h and the audio architecture overview. https://raw.githubusercontent.com/zeldaret/oot/main/include/audio.h and https://deepwiki.com/zeldaret/oot/6.1-audio-architecture | SEQ_NUM_CHANNELS 16; SequenceChannel holds layers[4]; SequencePlayer seqPlayers[4]; SequenceLayer sequenceLayers[64] preallocated; SEQTICKS_PER_BEAT 48. Establishes that OoT's thin textures are an AESTHETIC choice, not a hardware ceiling - the headroom was there and was not used.
CONTEXT (monographs, named for the depth pass) - Andrew Schartmann, 'Koji Kondo's Super Mario Bros. Soundtrack' (Bloomsbury 33 1/3, 2015); Tim Summers, 'The Legend of Zelda: Ocarina of Time: A Game Music Companion' (Intellect, 2021) and 'Understanding Video Game Music' (Cambridge, 2016); Karen Collins, 'Game Sound' (MIT, 2008) and 'In the Loop: Creativity and Constraint in 8-bit Video Game Audio', Twentieth-Century Music 4/2 (2007). | Collins's finding matters directly to the loop doctrine: loops were 'at least in part, an aesthetic that grew as the games became more popular and more complex' rather than a pure consequence of limited memory. These four are the acquisition targets if the grammar needs deepening beyond what is derivable from the open sources above.
NEGATIVE CONTROL / KNOWN GAP - No source located gives a measured distribution of LOOP LENGTHS IN SECONDS for the exemplar class; the only figures found are practitioner conventions (avoid loops under ~20s; ~30-40s typical) from community and blog sources, which are NOT of citable quality. FORM-05 is therefore stated in BARS (24-32 minimum, grounded on Zelda's Theme's 24-bar binary) rather than seconds, and any seconds-based gate the rig adopts must be derived by direct measurement of the named exemplar tracks, not asserted.
---
3. The measurable features
opening_interval_semitones: absolute semitone distance between the first two pitches of the lead line. GATE: <= 7. TARGET MODE: {2, 3, 4}. Exemplar distribution is fully enumerated in MELODY-01; a 6th/7th/octave opening is out of dialect.
head_cell_pitch_count: number of DISTINCT pitches in bars 1-2 of the lead. TARGET: 3 (accept 3-4). Zelda's Theme basic idea = a 3-pitch cell.
hook_onset_beats: beats from cue start to the first lead-melody note. GATE: <= 8 beats (2 bars in 4/4). Any longer is a standing intro.
melodic_pitch_alphabet: distinct pitch classes across the whole primary melodic statement. GATE for signature/diegetic/player-performable cues: <= 6 (target 5). Soft target for full themes: <= 8.
melodic_range_semitones: max minus min pitch of the primary statement. TARGET: <= 17 (octave + P4). Exemplar warp songs sit well inside this.
leap_fraction: proportion of consecutive melodic intervals larger than a M2. TARGET: 0.25-0.45, AND at least one leap >= P4 per 8-bar unit paired with a predominantly stepwise/neighbour answering half (MELODY-05).
contour_class_arch_rate: fraction of 8-bar units whose smoothed pitch trajectory classifies as rise-then-fall (or its inversion). TARGET: >= 0.60. Penalise exotic/novel contours - novelty must live in interval detail and harmony, not contour (Jakubowski et al.).
turning_point_gradient_unusualness: mean gradient between melodic turning points, z-scored against the exemplar-class centroid. TARGET: outside the common band while contour_class stays common - the earworm signature is common contour + unusual gradient.
scale_degree_5_dwell: proportion of melodic duration spent on scale-degree 5. TARGET: elevated vs a neutral-tonal baseline (named Kondo-ism).
tonic_triad_outline_score: fraction of bars 1-4 melodic pitches that are tonic-triad members. TARGET: >= 0.6.
phrase_grid_conformance: fraction of detected melodic segment boundaries landing on 2-bar boundaries. TARGET: >= 0.80. Units resolve to 8 bars.
sentence_detector: interval-sequence similarity between bars 1-2 and bars 3-4. GATE: >= 0.80 (literal or near-literal basic-idea repeat). Plus continuation check - mean motivic unit length in bars 5-8 <= half that of bars 1-2 (fragmentation).
cadence_type_at_loop_end: classify {PAC, IAC, HC, plagal, bVII-I, bVI-bVII-I, modal/weakened}. HARD FAIL: PAC at the loop wrap. REQUIRE one of {HC, plagal, bVII-I, weakened-chromatic}.
final_leading_tone_present: boolean - is raised scale-degree 7 sounding in the cadential chord at the loop wrap. TARGET: false for the weakened-cadence class.
picardy_third_present: boolean - minor/modal cue ending on a major tonic triad. TARGET: true for minor and modal cues.
loop_seam_smoothness: Medina-Gray five-axis seam score (meter, timbre, pitch, volume, abruptness) computed at the wrap point and at every modular phrase join. TARGET: smooth by default, with disjunction only where deliberately designed.
bar_repetition_ratio: fraction of one-bar clips that are exact repeats of an earlier clip (Hunt's method - identical note number, onset, duration). TARGET BAND: 0.55-0.70, centred on the 0.6279 corpus mean (SD 0.0349). Outside the band in either direction is a defect.
motif_count_per_loop: number of motifs whose pairwise similarity is < 0.5. GATE: >= 3. TARGET: 4 (Kondo's own stated figure for the SMB Ground Theme).
motif_contrast_index: mean pairwise dissimilarity across the loop's motifs. GATE: >= 0.5 - motifs must not sound like each other, which is the whole mechanism.
section_order_symmetry_flag: raise a flag on pure AABB / AAAA / single-section orderings. REQUIRE either an asymmetric order or a returning-material device (e.g. intro material reused as an internal cadence).
loop_length_bars: total bars before the wrap. GATE: >= 24 for a themed cue (Zelda's Theme = 24-bar binary). Seconds-based gates must be derived by direct measurement of the exemplar tracks - see the negative control in the evidence list.
harmonic_rhythm_chords_per_bar: TARGET 1.0 in the theme statement (accept 0.5-2.0); 0.25-0.5 for ambient/water/celestial cues.
tonic_prolongation_first4: boolean - are bars 1-4 confined to a tonic field (I<->IV or i<->IV) over a sustained common bass tone.
mixture_chord_rate: fraction of chords borrowed from the parallel mode. TARGET: >= 0.15, and at least one member of {bIII, bVI, bVII, iv, iio, Neapolitan} must appear per loop.
mario_cadence_present: detect the bVI-bVII-I succession. REQUIRE at every triumph/fanfare/reward/arrival cue; forbid its overuse elsewhere (cap ~1 occurrence per non-fanfare loop).
bVII_for_V_substitution_count: count of cadential approaches using bVII in place of V. Presence is a positive style signal.
secondary_function_count: count of distinct secondary dominants/leading-tone chords per loop (V/ii, V/IV, V/V, V/vi, viio/x). TARGET: >= 1 in any cue longer than 16 bars; purely diatonic loops fail the dialect check.
declared_mode_and_characteristic_degree_rate: each cue declares one mode; measure occurrences of the characteristic degree (raised 4 for Lydian, lowered 7 for Mixolydian, natural 6 for Dorian) per 8 bars. GATE: >= 1, and for Lydian the raised 4 must be approached by step (passing/neighbour), not leapt to.
bass_semitone_motion_rate: fraction of bass note-to-note moves of exactly +/-1 semitone. TARGET: >= 0.25 across the approach to any cadence; detect sustained pedal followed by chromatic descent as a named pattern.
voice_leading_displacement: mean total semitone displacement summed across voices between consecutive chords. TARGET: <= 3 (parsimonious). Chromatic substitutes must be justified by this metric, not by function.
extended_tertian_rate: fraction of chords carrying add9 / 11 / maj7#11 / 13. TARGET: elevated ONLY in ambient/water/celestial cues; near zero in up-tempo locomotion cues.
simultaneous_voice_count: median number of concurrently sounding independent parts (parts, not samples or layers). TARGET: 3-5. FLAG: > 6 in any signature theme.
voice_role_distinctness: classify each part as lead / bass-counterline / harmonic-ostinato / percussion. FLAG: any part unison-doubling the lead for > 20% of bars; FLAG: a bass that is pure root-doubling with no independent contour.
chord_voicing_openness: mean interval between adjacent chord tones in the harmony parts. TARGET: open (no close-position stacks) when timbres are harmonically rich.
rest_density_melody: fraction of beats in the lead part with no sounding note. GATE: >= 0.15 - silence is a compositional element, not dead air.
cross_rhythm_index: subdivision ratio of the percussion layer (triplet/swing) against the melody+bass layer (straight). REQUIRE a mismatch in the up-tempo/locomotion class - this is Kondo's own named source of forward motion.
syncopation_index: off-beat onset weight in the lead (Longuet-Higgins/Toussaint-style score), z-scored against the exemplar centroid. TARGET: within the exemplar band, not below it.
clave_cell_presence: detect a 3+3+2-style dotted cell (two dotted quarters plus a quarter) in the percussion or comping layer, sounding against straight melodic values.
tempo_bpm: log and compare against the exemplar-class distribution; earworm research links faster tempi to involuntary recall, so slow-by-default is a flag for cues meant to be carried.
locomotion_lock: measure the cue's primary pulse against the player-character's step/jump cadence in the target chapter. GATE: the pulse must be an integer or simple ratio of the locomotion rate - music not derived from the game's own rhythm fails Kondo's first principle.
singability_gate: can the primary statement be reproduced monophonically inside one octave using <= 6 pitches at a human-comfortable tessitura. Boolean, computed on the extracted lead - the hard hummability test, replacing the failed proxy.
region_identity_separability: classification accuracy of a held-out classifier assigning a cue to its region/culture from timbre + mode + rhythmic cell alone. TARGET: high. A cue that could belong to two regions fails Kondo's 'unambiguous sonic image' principle.
comparator_distance_vector: per-feature distance from the exemplar-class centroid built over the named track set (SMB Ground Theme, SMB3 Overworld, SMW Athletic, SM64 Main Theme, SM64 Dire Dire Docks, LoZ Overworld, ALttP Hyrule Castle, ALttP Kakariko Village, OoT Title, Zelda's Lullaby, Saria's Song, Song of Storms, Great Fairy's Fountain, Hyrule Field, Majora's Mask Astral Observatory, Song of Healing, Wind Waker Grandma's Theme). REPORT PER FEATURE, never as a single scalar - the collapse to one score is what let 14 unanchored themes ship.
Generated by harness/site/structure_site.py — the URL path is the repo path. review root