Imported from jtydhr88/japanese-composition-skills (plugins/japanese-composition/skills/jp-melody-writing/SKILL.md). Install upstream with npx skills add jtydhr88/japanese-composition-skills --skill jp-melody-writing. Copyright stays with the author.

Japanese-Style Melody Writing (Melodic Method)

Source: the complete system of Hiroshi Takayama, Popyurā Ongaku Sakkyoku no Tame no Senritsu-hō, Expanded Edition (ポピュラー音楽作曲のための旋律法 増補版, Rittor Music). This file reorganizes the book into a practical reference by compositional task. The large lookup tables (seven types of non-chord tones, motive-construction table, eight degrees of cadential openness, form formula collection, situation-based reverse-lookup skeleton, etc.) are in reference.md in this directory.

0. Core Concept: A Melody = One Tone Moving Through Time

• A melody is not "multiple tones arranged in space" but one tone moving through time (imagine a floating ball moving up and down).
• Listeners connect earlier and later tones by memory and retroactively update each tone's role (starting point → passing point → arrival point) — only upon hearing the 3rd tone is the 1st tone reinterpreted as "the start of an ascent."
• Every tone sits in a force field (場, ba) made of four forces:
  1. Gravity of register: high notes are tense — moving upward requires effort; low notes are settled — moving downward feels like releasing tension;
  2. Force of the mode: in-scale tones are stable, out-of-scale tones are pulled back;
  3. Force of harmony: consonance with the chord is stable, conflict makes the tone want to move — this is precisely the source of forward momentum;
  4. Force of beat and rhythm pattern: on-grid positions are stable, positions off the rhythmic grid are unstable.
• Two additional sources of expressive motion:
  • Everyday-experience associations: speech intonation (イントネーション), the physical load of bodily movement — stepwise motion = small gestures, hence smooth; large leaps = big gestures, hence effortful; repeated low notes = the strong vibration of a large instrument, hence a sense of heat.
  • In-song experience (designable by the composer): an unusual use of a tone can be accepted empathetically if it is prepared in advance within the song (letting listeners accumulate the relevant experience first); an ordinary use of a tone can become fresh through contextual relationships. This is the key to resolving the "banal vs. self-indulgent" dilemma.
• Segmentation (分節, bunsetsu): the same string of tones, segmented differently, = a different melodic experience. When writing a melody you must constantly anticipate "where and how the listener will segment." The units cut out by segmentation are phrases (フレーズ).
• Hierarchization of motion: tone → motion → phrase → motion-of-⟨tonal motions⟩ → large-scale structure.
  • Stacking and enlarging tension (tension↑ → more tension↑ → expansion↑) creates sustained expression and large forms;
  • The final release acts on everything that came before, producing an intense sense of relief (安堵感).
• Terminology convention (strictly distinguished in the book): pitch = the height of a tone itself; interval = the distance between two tones.

1. Expressive Rules of Pitch Motion

Direction (向き)

• For phrases of three or more tones, overall direction is judged by the relative height of the first and last tones; if the local motion arriving at the final tone opposes the overall direction, the sense of overall direction is weakened.
• Segment long phrases first, judge direction section by section, then assess the composite effect.
• The harmonic background can rewrite the meaning of direction: if an ascent lands on a stable tone by harmonic mechanics (e.g., G7→C landing on the tonic), the "questioning feel" is replaced by a "sense of resolution." To emphasize question/tossing-out, remove the elements that block it (don't land on a stable tone).
• A slow ascent with restrained volume and tension can express "longing gazing into the distance, pure yearning" rather than tension.
• Bodily analogy: ascent = stretching up to look far away (伸び上がり), descent = bending down to look closely (かがみ込む).

Width (幅)

• Stepwise motion (順次進行, junji shinkō):
  • Smooth and flowing, without obstruction; ascending/descending along the scale = a smooth increase/decrease of tension, extremely common;
  • Adding auxiliary tones a 2nd above/below the existing tones as ornamentation = increasing motion while keeping the contour, more florid.
• Semitones:
  • The narrowest interval, expressing subtle nuance and delicate tenderness (デリケート); a melody containing semitones = voluptuous beauty (艶やかさ) and depth;
  • Overuse (especially chromatic semitones with accidentals) → neurotic, morbid, grotesque (グロテスク) — must be restrained;
  • "How to use semitones well" is a major requirement of melodic sensibility (one of the two great sources of melodicity, alongside pentatonic density).
• Leaps (跳躍進行, chōyaku shinkō):
  • Dramatic, large emotional swings; the position of the leap is emphasized (whichever word the leap lands on, that word stands out);
  • A single leap should stay within a 6th; a perfect octave, however, is actually easy to sing and easy to hear, and is usable;
  • Too many leaps still make a melody hard to sing and unnatural — use with restraint.
• Repeated notes (same pitch):
  • Low register = calm confiding, understated narration (or forcible suppression); high register = tension continuously sustained, ceaseless shouting;
  • When pitch stays still, the listener's attention shifts to rhythm → combine syncopation and varied durations = percussive / speech-like (語り, katari — narration) expression.
• Width × speed:
  • The narrower the interval, the easier to sing and the more it suits any tempo; the wider, the more it needs slow tempo and preparation;
  • A large leap needs sufficient duration at its start or end (or both) to sound natural and singable; rapid continuous alternation between high and low notes stiffens most easily.

2. Preparing and Landing Leaps

• Upward leaps are easy, downward leaps are hard (moving while adding force is controllable; motions that release force are hard to stabilize).
• Two remedies for a wide downward leap:
  1. The landing point naturally carries an accent (アクセント) — anticipate it and use it deliberately;
  2. Lengthen the starting tone fully + diminuendo (ディミヌエンド) — build up emotion and tension at the start before leaping.
• Among narrow leaps, downward is actually easier (force release) — a "short high note leaping down" at the start of a phrase to create a phrase-head accent is a common device (the 「いつでも」「どこでも」 type, matching word intonation).
• Tame (タメ, windup/holding back): before a leap, place 1 to several tones moving in the opposite direction (usually stepwise), like winding up before a big gesture; it can also be placed mid-leap.
• Josō (助走, run-up): before a leap, place 1 to several stepwise tones in the same direction; the run-up tones may lie inside or outside the leap interval.
• Landing cushion: after arriving at the leap, follow with 1 to several stepwise tones in the same direction, softening the landing impact — especially effective for downward leaps.
• Rebound (reaction): after a large leap, stepwise motion in the opposite direction = a landing rebound; that rebound step can double as the tame for the next leap, forming a chain of "leap → rebound(windup) → leap → rebound(windup) → leap."
• Splitting a 7th: a 7th leap is most commonly split with an intermediary tone into 3rd+5th or 4th+4th (stepping-stone style, 飛び石, in two steps).
• Hard limit on same-direction consecutive leaps: leaps of a 4th or larger may continue in the same direction at most 2 times, and the total interval between the two ends must stay within 1 octave.
  • Compliant examples: 3rd+5th = 7th, 4th+4th = 7th;
  • Violations: 5th+5th = 9th, 4th+4th+4th = 10th — unnatural and hard to sing.
• A melody with many large leaps tends to sound instrumental and mechanical; but a large leap at slow tempo with unhurried preparation can draw an expansive, open (伸び伸び) melodic line.
• Leaps larger than an octave strongly sever the melody (they cannot be segmented as one phrase); reverse use: separate registers widely at segmentation boundaries → a pseudo-two-voice melody (a low-register motive and a high-register motive in dialogue).
• The value of "unnatural" motion (all require restraint — overuse kills the effect):
  • An unprepared large leap = a sudden emotional outburst (filling in intermediary tones makes it smooth but destroys the explosive force);
  • Not resolving a leap in the opposite direction but continuing the same way = a special sense of setsunai (切ない, aching bittersweet heartache);
  • Consecutive leaps with thin internal connection = disembodied lightness, an inorganic pop feel;
  • Displaced semitones with no perceptible scale affiliation = artificiality, detached irony;
  • Augmented intervals (増音程, zō-ontei) = a non-Western feel.
  • Master formula: embed local unnaturalness as a "hook" (引っかかり), then finally resolve it in a natural way.

3. Range Setting and the Human Voice

Basic Rules of Range

• First set the melody's central range at roughly 1 octave, and place the center of each phrase within it — otherwise the melody becomes diffuse, or even sounds like a mix of multiple voices.
• The whole song's actual range is about 1.5 octaves, at most 2 octaves (which happens to be the usable range of most voices and instruments).
  • Contained within 1 octave = a narrow-interval melody, intimate and easy to sing;
  • Spanning 2 octaves = strongly dynamic but hard to sing.
• Effects of a phrase's position within the range:
  • Center shifted high = bright, full of tension; a tone breaking above the range ceiling spikes in tension; extreme highs = an anguished (苦しげ) expression;
  • Center shifted low = calm and relaxed; falling below the floor = heavy (重々しい); extreme lows = powerless slackness.
• The same motion has different effects at different positions in the range:
  • An ascent in the middle of the range = moderate tension; an ascent toward the top, breaking out of the range = tension conspicuously amplified;
  • An ascent climbing from a low position back to the middle = the tension increase is offset by "low notes turning soft," and may even act as relaxation;
  • The same phrase-final ascent: low register = a light-hearted question; high register = an earnest plea, even an interrogation (詰問).
• Pitch character is relative: the same middle C is heard as a "high note" in a low-register melody and a "low note" in a high-register melody.

The Human Voice (声の音域)

• Male voices are about 1 octave lower than female voices; by convention male voice parts are notated 1 octave higher than sounding. In pop practice: male voices rarely use the low register and often use the high register; female voices rarely use the extreme high register and commonly use the low register.
• Chest voice (チェストボイス, ≈ jigoe/natural voice):
  • The root of pop-music reality (リアリティ) and intimacy; the character of pitch motion is expressed most intensely here;
  • The strained quality of the high register can express earnestness (the urgency of wanting to convey something).
• Middle voice (ミドルボイス):
  • Half-closed vocal folds, extending the range upward; the timbre is bright and open = a positive, projecting strong expression;
  • The same high phrase: sung in chest voice = strained and imploring (苦しげ); sung in middle voice = expansive, gazing into the distance — register choice changes the melody's character;
  • Whether a male singer has a middle voice, and where his passaggio (換声点, kanseiten — register break) lies, is a major precondition determining range and key (キー) — confirm it first.
• Falsetto (ファルセット/裏声):
  • Treat it as a different instrument from chest voice: pure and delicate, suited to refined, tender (デリケート) expression;
  • Limitations: few overtones → diction slightly blurred; dynamics hard to vary → unsuited to intimate confiding and large emotional swings;
  • Flipping into falsetto from chest voice = the desperation (懸命さ) of reaching a high note even at the cost of the voice breaking → intense longing, ultimate setsunasa.
• For singers with an awkward passaggio: design registers phrase by phrase (this phrase in chest voice, that one in falsetto/middle voice), so no single phrase crosses the break.
• Set the melody's basic octave in the middle of the usable extreme range — leaving headroom while keeping the option of deliberately touching the limits for expression (a strained high note conveying depth of feeling).
• The shōwa (唱和, inner sing-along) effect:
  • Listeners hum along with the melody in their heads — after being induced into entrainment by accessible repetition, a subsequent high note still feels "high and tense" to the listener even if the singer produces it effortlessly;
  • A sudden high note at the song's opening or a phrase's start gives no time to entrain → it is externalized and appreciated as display (a sense of virtuosity);
  • When writing, be explicit about whether a passage will be "sung along with in entrainment" or "appreciated from outside," and actively induce the intended mode.

4. Duration and Cadence

Usable Range of Note Values

• Common note values: 16th note to whole note; tuplets are normally only 1-beat or 2-beat triplets (32nd notes, quintuplets, etc. appear only occasionally to cram in lyrics).
• Principle: pop music must let listeners stay entrained — note-value combinations must not make the listener lose the beat.
• For a whole song in triplet feel, don't write triplets everywhere; renotate in compound meters 6/8, 9/8, 12/8.
• First examine character at a moderate tempo around BPM ≈ 100, where "note value ≈ note length," then consider the shifts at extreme tempos.

Long Notes and Short Notes

• Long notes = sustained force = tension; expression can be shaped inside the note (vibrato ビブラート, crescendo/decrescendo) → climaxes commonly place a long note in the high register (the maximum-intensity expression of tension).
• Consecutive long notes + little pitch motion = sustained tension; but sustaining too long turns slack and boring — insert leaps, or weave short notes between the long ones to revive it.
• Long notes + large leaps combined = smooths out large pitch motion while amplifying the sense of scale (スケールの大きさ).
• Each long note carries strong expression: sung forte = power and conviction; sung piano = gradually fading, hanging by a breath.
• Short notes = light, agile, lively; sung forcefully = sharp, brusquely flinging away (突き放す). Short notes in the low register are offset by heaviness (重い).
• Short notes + wide leaps are hard to sing and come with built-in dynamic swells — usable intentionally as accent/load; a large leap still needs "motion-preparation time" before it.
• Narrow intervals × short notes = melody approaching speech (語り/セリフ) (news-anchor speech rate ≈ 8th notes at BPM 120, which does not sound fast).
• Continuous short notes across a wide range = light rolling motion; kept up too long = neurotic, obsessive.
• Mixing long and short: long notes = principal (center of gravity / natural accent), short notes = subordinate — regardless of beat position.
  • Common cycle: short notes approach (アプローチ) a high long note → tension → a low short note releases → approach again…;
  • Extreme compression = alternating long-short-long-short skip feel (スキップ) (common in 6/8, 12/8) — handy but prone to yotei-chōwa (予定調和, predictable resolution / cliché predictability); beware banality.

Duration Treatment at Cadences

• Lengthening the cadence tone = motion stops; two follow-ups: add one short note to release the tension, or use a rest for de-tensioning (脱力).
• Duration alone is not enough to close a cadence: if harmony and mode (scale degree) do not cadence at the same time, the sense of ending remains half-finished — usable in reverse: let the chorus settle while the whole song stays unresolved, leaving a "words left unsaid" feeling.
• Large cadences (section-final / song-final):
  • Ritardando + decrescendo bleed off energy; a "notated ritardando" = writing each note before the cadence progressively longer, or inserting one bar of 2/4;
  • Variant: stretch the note before the cadence fully (tame) then end on a short note — tension held to the last instant, then released in a flash;
  • Further: the melody omits the cadence tone, handing it to the accompaniment — a dreamlike afterglow like breath cut short;
  • Reverse: no slowing, no lengthening, close with an accented short note = a decisive, clean, strong will (can be reinforced with an additional accented short note).
• Rests are part of the melody:
  • A powerful segmentation tool — at the end of a long phrase, don't fill the full value with a long note; place a rest to make the segmentation explicit yourself (or use a breath mark ブレス or a slur スラー to frame the phrase);
  • A short rest inside a phrase = de-tensioning, functioning like a low short note;
  • Wherever the lyric contains a sokuon (っ, geminate stop), a short rest necessarily results.

5. Beat Position and Syncopation

Beat Position of Phrase Openings and Closings

• A song opening with no accompaniment + flat note values easily falls into a meterless state (beats cannot be grouped) — beware; it can also be used deliberately as an unstable, mysterious introduction.
• Mid-song meter changes: a section-level change = dramatic but risks splitting into two songs (keep unity with the same melodic figure); locally inserting one bar of 2/4 = a tame before the chorus that launches a pickup — a stock device; at any meter change, anticipate what the drums and bass will do.

Syncopation (シンコペーション) and Anticipation (アンティシペーション)

• Syncopation = a weak-beat (or offbeat) note tied into the following strong beat → that note becomes "strong-beat-ized" → the next step is dragged later = a tame feel, unhurried lag.
• Anticipation = a strong-beat note played early on the preceding weak beat → the progression arrives early = urgency, restlessness, a driving rush.
• Precondition for both: the listener already clearly perceives the beat — used a cappella or at the song's opening where the beat is not yet established, they are not only ineffective but disrupt the sense of meter.
• A melody glued to the beat throughout = plain, earnest and forceful, but stiffly foursquare (四角四面), and since the accompaniment strikes on the beat, the melody doesn't stand out; moderately detaching from the beat (offset the opening + timely syncopation) → the melody floats up off the beat, free and expansive (伸び伸び).

Coordinating with the Rhythm Pattern (リズムパターン)

• Write the melody with a rhythm pattern already implied inside it: while composing, play back an 8-beat/16-beat groove in your head and let the melodic motion move within that groove (ノリ) — avoid "forcing a melody into a ready-made rhythm pattern."
• Three relationships with the backbeat (the snare on beats 2 and 4 in an 8-beat):
  1. Have the melody rest at the backbeat positions → a call-and-response (掛け合い) with the backbeat;
  2. It need not be strictly vacated; consciously yielding to it is already effective;
  3. Reverse: melodic strong accents in unison (ユニゾン) with the backbeat — you can partially overlap and let the backbeat show through where the melody rests.
• 8-beat melodies often move on the offbeats (ウラ) of quarter notes; 16-beat melodies on the offbeats of 8th notes → syncopation/anticipation arises naturally.
• Ura-nori / omote-nori (ウラノリ／オモテノリ): a groove consciousness taking the offbeat/onbeat as the starting point of each gesture (start → resolution); neither is superior, but the feel differs:
  • Omote-nori = grounded steadiness, striking power; ura-nori = bouncing, flowing lightness;
  • The groove determines which words get focus (in ura-nori, words starting on offbeats become focal) — to emphasize a word, adjust note placement according to the groove.
• Bounce (バウンス): offbeat notes shift later, closer to the next onbeat (≒ onbeat notes lengthen); the slower the tempo, the bigger the bounce; fixed at triplet values = shuffle (シャッフル); in a bounced song the melody can use triplets without incongruity.
• The rhythm pattern belongs to the accompaniment: leave gaps in the melody so the pattern shows through the gaps (a three-dimensional sound); don't trace/duplicate the rhythm pattern with the melody — redundant, cluttered, mutually canceling.

The Motion of Rhythm (リズム) Itself

• Rhythm = grouped note values + accent placement (different accent placement = a different rhythm); high notes naturally act as accents — you can build rhythm by placing high notes alone; swapping on/off positions (ウラオモテ) radically changes a rhythm's character.
• Plain repetition: stable but rigid; enlarging the repeated unit to 2+ bars reduces the sense of formality; the fewer note-value types, the stronger the repetition effect and the heavier the formality.
• Displaced repetition: repeating a 1.5-beat (dotted quarter) unit in 4/4; repeating a 2-beat unit in 3/4 = hemiola (ヘミオラ) — the displacement generates tension.
• Deformed repetition: overall compression = impatient, pressingly earnest; overall expansion = unhurried and settled; lengthening via anticipation = paradoxically more urgent.
• Partial repetition: leisurely echoing the phrase ending after a gap = echo (エコー) — introspection, lingering regret (心残り), murmuring to oneself; progressively shortening and beat-compressing the repeated part = tatamikakeru (たたみかける, pressing in step by step) — urgency.
• Contrast: rhythmic shapes must have vivid characters to be recognized; the most legible = short-value groups vs. long-value groups; a latent contrast with the order flipped prevents boredom without seeming arbitrary.
• Rhythm with unclear segmentation = long-winded; segmentation that is too self-evident = dull — accumulate tension with "hooked" note values and beat positions, then release at a plain-spoken position (a sense of opening, 開放感).

6. Modes and Pentatonic Scales

Major Mode / Minor Mode (長旋法／短旋法)

• A mode = selecting tones from the 12 + assigning roles (tonic, etc.); the key is determined by the tonic's position, not by the mode type; the same set of tones with a different tonic = a different mode (all white keys can be C major or A minor).
• Major-mode mechanics: the leading tone (導音, dō-on, degree vii) resolving up a semitone to the tonic is strongest > dominant (v) > subdominant (iv);
  • Mechanics ≠ compulsion — the leading tone moving by arpeggio into another chord tone is also very common; the major mode overall has few constraints and free motion;
  • An overly strong functional feel (leading tone → tonic, tracing dominant-to-tonic harmony) = prim and stiff (school-anthem feel / punk's blunt shouting);
  • Major key ≠ cheerful: graduation songs and other works of sorrow and nostalgia are mostly in major.
• The three forms of the minor mode:
  • Natural minor scale: no leading tone; motion is free but melodies easily go flat and featureless;
  • Harmonic minor scale: vii raised a semitone as leading tone (extremely strong pull to the tonic); an augmented 2nd arises between vi–vii — avoid connecting them directly;
  • Melodic minor scale: ascending, both vi and vii are raised (eliminating the augmented 2nd; the raised vi strongly tends toward the leading tone); when vii is not acting as leading tone it is not raised and moves toward vi (descending form).
  • Standard practice: proceed freely in natural minor, and locally mix in harmonic (or melodic) minor at section boundaries to "punctuate" — strong effect but easily predicted and contrived-sounding (わざとらしい); use sparingly.
• Beyond the leading tone, minor has several other semitone locations → well suited to emotional (情緒的) melodies; minor key ≠ dark (much hard rock and many jazz standards are in minor).
• Church modes = grab the characteristic tone (each differs from the major/minor mode by exactly 1 tone; full table in reference.md):
  • Dorian (ドリア) = major 6th; Phrygian (フリギア) = minor 2nd; Lydian (リディア) = augmented 4th; Mixolydian (ミクソリディア) = minor 7th;
  • Place the characteristic tone at the right moments in the melody + support it with chords containing the characteristic tone; never using the characteristic tone = the mode is imperceptible, wasted;
  • Discrimination: a melody with lowered vii is usually "major mode + blues pentatonic," not Mixolydian.

Pentatonic Scales (ペンタトニック) — a Major Source of J-Pop Melodicity

• Pentatonic scales have no leading tone and cannot form a dominant chord → cannot establish a key by themselves; the mechanical core = "neighbor-tone tendency":
  • Tones tend to move to adjacent scale tones (adjacency comes in 2nds and 3rds; 2nds are easier); leaping to a farther tone requires force, and after the leap the tendency is to reverse direction.
• Four typical figures ("pentatonic-sounding"):
  1. Move to a neighbor and fold back;
  2. Move to a neighbor and continue in the same direction;
  3. First "hook" (引っかける) to the neighbor in the opposite direction, then leap back two scale positions (= the pentatonic version of tame; the variant whose first step is a 2nd is the most common);
  4. Overshoot past the target neighbor to the far side, then fold back.
• Three figures that destroy the pentatonic feel:
  1. Frequently leaping to tones 3+ scale positions away (one such leap followed by a return to normal is fine);
  2. Leaping 2+ positions and then continuing in the same direction;
  3. Consecutive leaps or repeated same-direction leaps.
  • Of these, arpeggio-like (broken-chord) motion is the most destructive — but arpeggio leaps across phrase boundaries are fine.
• Same-direction stepwise motion ≤ 3 steps: 4 steps sweeps the entire scale = a characterless scale-run that merely announces "I am pentatonic."
• Mode shifting (転旋法, ten-senpō):
  • Two pentatonics a fifth apart differ by only 1 tone → easy to switch, commonly used, enriching the pentatonic's small tone supply;
  • Switching between major/minor mode ⇄ pentatonic (usually without a key-signature change): a pentatonic opening (intimate, singable) → splice into stepwise scale motion (semitone momentum) as complements; locally mixing the leading tone into a pentatonic to fit the harmony — freely adjusting distance between "pentatonicism" and "functional harmony."
• The pentatonic tonic may differ from the key tonic (e.g., C major with an E-minor pentatonic = floating, modern color) — a standard way to avoid pentatonic flatness; any major/minor pentatonic built on a scale degree of the key is usable (if it contains an out-of-key tone, drop that tone).
• Pentatonics are highly independent of harmony: tones that look like appoggiaturas/passing tones need no resolution and may leap away freely; the same harmony fits many pentatonic phrases and the same pentatonic phrase fits many progressions — "coexisting well with the harmony while acting independently and freely" is the essence of pentatonic melody.
• Weakness: only 5 tones invite cliché, and no semitones means emotional scarcity — you can write in fixed kobushi (こぶし, melismatic pitch inflection)-style semitone turns (overuse becomes garrulous).
• Diluting the pentatonic feel: insert large leaps or arpeggio-like motion. In practice, regulate "pentatonic density" phrase by phrase (labeling each phrase as ~pentatonic (strong)/(weak)/intermediate); even when not using a pentatonic, always ask how pentatonic the current melody feels.
• Pentatonic cadences: the tonic = the strongest cadence tone, dominant next; arriving via a pentatonic neighbor or an ascending 4th is stronger; the precondition for a strong cadence = the pentatonic tonic coincides with the key tonic; otherwise only soft cadences are available.

Blue Notes (ブルーノート)

• Definition: temporary semitone lowerings of degrees iii, v, vii in major (the "melancholy tones") — not a fixed change to minor-scale tones, but:
  • Lowered temporarily, or deliberately sung flat with the "true tone" as premise;
  • The typical usage is a before/after contrast of "original tone → blue note," or ornamental oscillation between the two.
• The effect is strongest when the harmony does not contain the blue note (the semitone collides with the chord's diatonic 3rd): blue iii against the 3rd of I, blue vii against the 3rd of V — a head-on collision, the fiercest blues color.
• Often packaged as the blue-note pentatonic (ブルーノートペンタトニック) (a minor-pentatonic shape containing ♭iii and ♭vii) used over major; blue v does not join the pentatonic and almost always resolves by semitone step up/down (used as a passing/ornamental tone).
• To "dye" the whole song blue: add chords such as ♭III, ♭VI, ♭VII, Vm, I7, IV7 (usable even when the melody has no blue notes; this is a non-functional usage carrying no harmonic function).

7. The Relationship Between Harmony and Melody

Chord Tones and Non-Chord Tones

• Chord tones (和音内音): stable, straightforward, plain; but easily buried in the chord, and inevitably leap-heavy, lacking smoothness. Stability ranking: root > 5th > 3rd.
  • A purely chord-tone melody easily runs parallel to the harmony and gets swallowed by it; but when the harmony moves frequently or contains local modulations, it can actually highlight the harmonic interest.
• Non-chord tones (和音外音): unstable, tense, hard to bury; their urge to resolve (basic form = moving by 2nd up or down to a chord tone) is the greatest engine of melodic momentum.
  • Tension control: non-chord tones a semitone from a chord tone are strongest (hard and fierce, to be used with care); the longer the value and the stronger the beat, the greater the tension.
  • In pentatonic/blues contexts, modal force takes priority and non-chord tones need not resolve.
  • Non-chord tones may be used consecutively; other chord tones may be interposed mid-resolution; unresolved / irregularly resolved non-chord tones = subtle feelings, an ineffable afterglow, inorganic expression (also a resource, but too long or too strong harms the sense of harmony).
• Definitions and resolutions of the seven types of non-chord tones (passing tone / neighbor tone / appoggiatura / anticipation / escape tone / suspension / suspension-appoggiatura) → reference.md Table 1.
• The more chord tones a chord has (triad → 7th → 9th → 11th → 13th) → the fewer non-chord tones remain → the freer the usable tones, but the blander each tone's relationship to the chord, losing the delicious mechanics of non-chord tones.
  • The chord-tone/non-chord-tone status of the same melody changes with the chord symbol (under C / CMaj7 / CMaj7(9), the identities of B and D differ) → writing a melody requires simultaneously imagining the chord sounding with it;
  • The whole song needs a plan for "how many chord tones, and where and why the count changes" — a 9th chord suddenly appearing in a triad-dominated song = the forces at that spot lurch, unbalancing the structure.

Melody within Functional Harmony (機能和声)

• Function groups and progression rules (T→anything; D must return to T, D→S forbidden; S = buffer between T and D) → reference.md Table 2.
• Stock progressions at the instant of the chord change: leading tone → tonic (semitone up), degrees iv/vi → down a 2nd, dominant degree → hold or up a 4th — following them = a strong sectional/cadential feel; within the dominant chord itself, arpeggios and steps are free.
• The melody should avoid implying "forbidden progressions" (e.g., a melody strongly implying V→IV) — done deliberately, it yields a non-tonal, modal effect.
• Yotei-chōwa alert: a melody traveling with functional harmony (especially tones over the dominant chord) is extremely easy to predict — break textbook resolutions with suspensions, appoggiaturas, and anticipations.
• A section may cadence directly on the dominant chord, and the next section need not start from the tonic chord (a strong sectional break permits harmonic discontinuity, e.g., after a G7 section, restarting straight from F).

Melody within Non-Functional Harmony

• Modes without a leading tone (natural minor, pentatonic) often pair with non-functional harmony: chords planted directly against melody tones, parallel motion, harmony that doesn't progress and serves only as rhythmic accent, chains of 4ths (lacking the dominant→tonic thrust).
• Weak momentum → songs easily go flat; countermeasures: a combined tactic of rhythmic-density changes, register leaps, pentatonic ⇄ scale switching, appoggiatura resolutions, and locally borrowing V7→I to punctuate sections.

Mutual Independence of Melody and Harmony

• Melodic motion and harmonic motion need not coincide: melody stalls while harmony moves, they move in opposite directions, the same melody over different harmonies — all are sources of three-dimensional depth (exploit them actively; they are not errors).
• A melody that traces (なぞる) the harmony = flat and monotonous (occasionally useful for a unified sense of power).
• At least one of the bass (lowest voice) and the melody should leave the root: melody constantly coinciding with the root = stable but sonically impoverished; change the chord's bass note (inversion) or have the melody avoid the root.

Chords First or Melody First

• The trap of fixing a complex progression first: the melody degenerates into "walking along the harmony"; harmonic rhythm becomes formulaic (one chord per bar, dominant at the end of 8 bars); transplanting a genre's typical progression ≠ obtaining that genre (characteristic progressions exist bound to rhythm, melody, and position in the form).
• Practice path: freely invent melodies over 1–2 chords → over a 3–4-chord loop progression (play the loop until your body memorizes it, then "hold the harmony in mind while inventing melody") → once fluent, invent even complex progressions in relation to the melody.
• Conclusion: melody and harmony are parallel temporal structures that should be conceived simultaneously and interlocked tightly — neither takes precedence.

8. Fitting Lyrics and Melody (歌詞と旋律)

Intonation (イントネーション)

• Principle: the melody's rises, falls, and stresses match the word's pitch accent — otherwise meaning becomes hard to catch or is heard as a different word (「雨が晴れて」 "the rain clears" → 「飴が腫れて」 "the candy swells").
• Pitch accent follows standard Japanese (標準語) (standard ≠ Tokyo dialect; when unsure, consult an accent dictionary アクセント辞典).
• Deliberately defying the accent = a linguistic hook, expression of large emotion — but adjacent words must have correct accent to preserve the sentence's meaning; setting particles on high notes can build emotion in rising steps.
• Phrase endings: rising ending (語尾上行) = question, appeal, hope; falling ending (語尾下行) = answer, assertion, resignation.

The Rhythm of Words

• Chōon (長音, ー, long vowel): give it a long note value (or re-articulate for emphasis); stretching the wrong place changes meaning (「彼が好き」 "I like him" → 「カレーが好き」 "I like curry") — align the stretched note with the bunsetsu (word-phrase) boundary.
• Sokuon (促音, っ, geminate stop): express with a rest; or lengthen the preceding vowel, then cut it off short and strong (「はしーぃって」).
• Hatsuon (撥音, ん, moraic nasal): merge into the previous syllable, or give it a short, weak independent note; a long strong-beat 「ん」 is hard to sing and easily buried (unless you make it float by taking it off the chord tones).
• Devoiced vowels (無声音) (devoicing of u/i after t/k/f/s/p, e.g., the く of 「たかく」): no perceptible pitch — must get a short note, otherwise the phrase's musical meaning changes (the effect of figure repetition is lost).
• Consecutive identical vowels (「こと＋おもってる」): insert a rest between the two notes, or change pitch, to prevent them merging into one long tone that swallows the second vowel.

Segmentation and Emphasis

• When lyric bunsetsu and melodic segmentation coincide, comprehension is easy; misalignment changes meaning (「君が来るまで待ってる」 "I'll wait until you come" → 「君が車で待ってる」 "you're waiting in the car").
• Uses of deliberate misalignment: breaking before a particle to suspend interest; cutting mid-word to create tension (「わ・た・し」 syllable-by-syllable insistence); stretching the phrase head as tame to induce expectation — use only where word meaning survives.
• Letting the musical phrase override linguistic segmentation = a comprehension burden on the listener → reverse-used as the obsession and earnestness of "singing it out even so"; all-equal 8th notes + preserving word boundaries by pitch motion alone = a single-minded (一途) devotion.

Lyrics-First and Music-First (詞先・曲先)

• Shisen (詞先, lyrics-first): writing only the catchphrase at the head of the sabi (サビ, chorus/hook) lyrics-first is already effective; even merely reserving positions for sokuon words (「ずっと」「もっと」「きっと」) is effective. Give the words to be highlighted long values, high notes, and kime positions.
• Four repairs for kyokusen (曲先, music-first):
  1. More syllables than notes → split note values or add notes before/after;
  2. Fewer syllables than notes → merge note values;
  3. Counts match but rhythm doesn't fit → change values/rhythm;
  4. Contour conflicts with pitch accent → adjust pitches while minding segmentation.
• Don't finish the melody to the last detail in advance — keep room for revision, and iterate back and forth with the lyrics; take care not to destroy repetition/contrast structures to accommodate words — when necessary, go back and change the lyrics.

Ji-amari / Ji-tarazu (deliberate mismatches)

• Ji-amari (字余り, too many syllables): words crammed into short values, patter-like → a breathless desperation; compressing the notes at the phrase head = a forward-toppling surge; repeating that rhythm = step-by-step advancing earnestness. Overdone, it becomes strings of equal values, a speech-like melody; if the emphasized word earns no sympathy it slides into comedy.
• Ji-tarazu (字足らず, too few syllables): words stretched, articulated over multiple pitches → arrests attention, pouring in deep feeling; stretched too far = word meaning dissolves into scat (スキャット, meaningless vocalise).

Rhyme and Rhetoric

• Rhyme (韻): alliteration/end rhyme; the most natural setting keeps the same rhythm and beat position at rhyme points while varying pitch motion (rhyme automatically brings rhythmic neatness; varied pitch keeps flow and avoids stiffness); identical melodic figures at rhyme points = strong formality; rap = near-speech in non-rhyming parts + leaps in and out at the rhymes.
• Beware shichigo-chō (七五調, 7-5 meter), san-san-nana byōshi (三三七拍子, 3-3-7 clapping meter) and other Japanese stock rhythms slipping in unconsciously (over-familiar to Japanese listeners; check especially when writing rap).
• Parallelism (並置): sequences of like words/phrases → melodic rhythm repeats with pitch climbing stepwise to drive the climax.
• Inversion (倒置): be explicit about "which word this inversion emphasizes" → charge up with a rest-tame before the key word, or delay the key word by one bar; a catchphrase can be paired with unison (ユニゾン) accompaniment.
• The conjunction between lines steers the melodic trajectory:
  • そして/だから (additive/causal) → the melody continues and repeats in understated narration; or receives it grandly, lifting the register into a "cause-and-effect climax";
  • だけど (adversative) → change the melody's character, a large sectional break (narrow-range stasis vs. wide-range leaps in opposition); or connect blandly with the same character = an objective, bird's-eye view;
  • ところで (juxtaposition) → still keep some relationship between the two lines (same character = implying a bird's-eye view; rising vs. falling response = implying cause and effect).

9. From Motive to Phrase to Section: Principles of Construction

The Hierarchy of Terms

• Partial motive (部分動機) (cannot stand alone but has a distinguishable feature; ≈ a word; labeled a1, a2, with ' for variants)
• → Motive (動機, dōki / モティーフ): the smallest melodic unit that can stand alone, at least 1 bar (shorter, and its beat position cannot be judged independently); ≈ a bunsetsu; labeled a, b; on repetition it usually inherits the rhythmic profile and repeats with variation.
• → Small phrase (小楽節, shō-gakusetsu): usually 2 motives; labeled A1, A2.
• → Phrase/period (楽節, gakusetsu): most often 8 bars, 4 bars also common (other bar counts are mostly additions to or omissions from 8 or 4); a whole unified by one mood (ムード); labeled A, B.
• → Setsu (節): the strophe corresponding to verse 1 / verse 2 (strophic form, 有節形式).
• Different motives may share partial motives (e.g., a'' = a1' + b2) — an important means of structural unity.
• The character of identical material is determined by context: the same motive is a small climax after a static phrase, but a calming answer after a soaring phrase.

The Six Causes of Segmentation (区切り), by Clarity

1. Long values and rests (clearest);
2. Interval leaps (severing pitch continuity);
3. Repetition of a motion pattern (each repeated unit forms its own segment);
4. A clear cadence (a definite ending = strong segmentation);
5. A clear beginning (retroactive recognition: only when the next phrase's start is recognized is the previous segment confirmed as ended);
6. Repetition/contrast (likewise retroactively recognized).

• Segmentations of 2, 4, and 8 bars are easiest to grasp; odd-bar groupings (3+3+2, etc.) are hard to perceive.
• Strong segmentation (tonic + strong beat + tonic chord + long note/rest) = a sense of fresh start — good for opening a new section and reinforcing repetition/contrast; but it chops the song in half, so use cautiously anywhere except section ends.
• Weak segmentation = smooth continuity, sustained atmosphere; can blur structure or create a sense of stasis.

The Four Principles of Construction

1. Response (応答, ōtō): given two segments, the latter is heard as a reaction to the former (the most basic engine; need not be a literal Q&A).
   • Like phrases responding = echo type; contrasting phrases responding = continuation type; the "answer" may even come first (the stable form first, the question after);
   • Paired (ペア) responses are most stable but yotei-chōwa; juzu-tsunagi (数珠つなぎ, bead-chaining) chained extension develops richly but risks sprawl;
   • Responses run in parallel at multiple levels of motive/phrase (a motive responds both to the immediately preceding motive and to the whole previous phrase).
2. Repetition and contrast (反復と対照): a response of like character = repetition; of unlike character = contrast.
   • Each element is judged separately: rhythm repeated + pitch contrasted (or vice versa) is the most common intermediate state;
   • Repeat at most twice; from the 3rd consecutive occurrence of the same phrase, listener interest drains away — variation/development is mandatory;
   • Weak repetition (few shared elements, heavy variation, e.g., repeating only the rhythm pattern) escapes the "boring by the 3rd time" rule — even 4+ repetitions stay fresh;
   • Variation devices: slight pitch/rhythm changes, beat shifting, rhythmic augmentation, pitch transposition (sequence), interval expansion — an ascending sequence = stairstep exaltation; local ascent + macro-level descent = complex, subtle expression; gradually widening intervals = exaltation, narrowing = closure;
   • Conditions for contrast: the preceding phrase's character is vivid enough + the following phrase's character is opposite; multiple elements reversed simultaneously = strong contrast; different-but-similar characters = a contrast-free "unified mood" (usable in quiet sections);
   • Tension accumulated by multiple ascents can be resolved at once by one sufficiently large descent; deliberately answering with a small descent = energy not fully spent, passing lingering regret (心残り) on to what follows;
   • Summary: melody = a craft woven from repetition and contrast at various strengths — distribute the two to produce "continuation feel" and "sectional feel."
3. Recurrence (再現, saigen): repeating a phrase at a distance of a full phrase-unit or more — "the protagonist returns from a journey"; the same phrase seems changed in character after time has passed.
   • Preconditions: the original phrase was well memorized (repeated earlier) + the middle section is substantial ("the journey was full enough");
   • Carries the persuasive closure of fu ni ochiru (腑に落ちる, sudden deep understanding) → best placed at the end (A-x-x-A); recurrences elsewhere also tighten the material and give one phrase multiple meanings.
4. Continuation and shiritori (継続としりとり, word-chain linking): the last 1–several tones of a phrase reused as the next phrase's opening → smooth connection by the continuation principle, segmenting without severing.
   • The more tones shared, the more it sounds like repetition (shiritori at partial-motive scale is heard as repetition);
   • Weak adhesion — long sections built on shiritori alone sprawl; combine with repetition/contrast.

Setting the Climax (クライマックス)

• Position: usually at 2/3–3/4 of the segment, before the phrase's resolving tone (matching the golden ratio and ki-shō-ten-ketsu); the head-climax type also exists.
• Size scales with the containing unit, nesting layer by layer: a motive has its apex tone → a phrase's climax = one particular motive → a setsu's climax = one phrase, and that phrase is called the sabi; within the sabi, the climax again falls in the latter part per phrase-level rules.
• Half the effect of a climax depends on how it is wound down afterwards: when a climax feels thin, first revise the endings of the motives/phrases (add a descending close after the apex, magnify the phrase's resolution, rework the ending to carve a clear apex) — often that alone rescues the section.
• Purpose: preventing shiri-subomi (尻すぼみ, petering out / anticlimax) and flatness; even a sabi that opens high must be checked for a sufficient climax in its second half and a substantial close (avoiding a dragon's head with a snake's tail, 竜頭蛇尾).

The Principle of Section Count (full table in reference.md, Table 7)

• One-section construction = the repetition principle: [conclusion](−conclusion…); dance-music riffs, counting songs; repetition seeping into the unconscious = heightened corporeality (a repeated coda chant triggering sing-along).
• Two-section construction = the response principle (most basic): [introduction]−[conclusion]; climax in the latter section or just before it.
• Three-section construction, two types:
  • Type 1a a−a'−b (outset−development−resolution): the second section is a varied intensification of the first; stable;
  • Type 1b a−b−c (outset−another viewpoint−integrated resolution): this is the A−B−sabi construction most common in Japanese pop;
  • Type 2 a−b−a' (ternary form) = the recurrence principle: the climax is in the middle section b; b must be more emotional, differing in mood from its surroundings; the final section is usually varied to gain finality.
• Four or more sections = compounds of two/three-section forms; using all-different motives risks sprawl → share partial motives; expanding three into four = repeat the first or last section (repeating the middle is rare).
• Temporal principles: ki-shō-ten-ketsu (起承転結) (ki = introduction, shō = repetition and stabilization, ten = large contrast and climax, ketsu = conclusion) and jo-ha-kyū (序破急) (slow intro → a climax section that articulates time → crisp close) — both are typical paradigms, not absolute rules.
• The two master questions of construction design: how many sections, and in which one the climax sits; through what ladder of characters the climax is reached. Universal device: give weight to the opening (出だし) of each phrase, and use it as the material of contrast.

10. Song Forms and Whole-Song Assembly

The Two Great Setsu Forms

• Verse–chorus form (ヴァース＝コーラス形式 / メロ＝サビ形式):
  • Verse (ヴァース / hira-uta 平歌 / A-melo Aメロ) = introduction and narration: typically narrow range + short values, katari (speech-like); a distinctive hooky figure can also open;
  • Chorus (コーラス / sabi サビ) = the setsu's climax: the catchphrase type of short repeated motives (common in rock) or the long-lined utai-age (soaring delivery) type (the Japanese-style sabi, ballads);
  • Pillars = the contrast principle (the gap between verse and chorus = explosive release) + internal repetition;
  • Extensions: A−A' −B (repeated verse); A−(bridge ブリッジ)−C (the short Western connective section, raising register while heating up); A−B (B-melo Bメロ)−C (sabi) (the J-Pop staple: a second verse of different character); dance type A×n−B.
• Ternary form: A (melo)−B (sabi)−A (A') — sabi in the middle, final section a recurrence; B unfolds A's content more emotionally / from a wider vantage; returning to A = deep realization and feeling; extension A−A'−B−A'' = textbook ki-shō-ten-ketsu.
• Patterns of expansion from A into sabi: narrow-interval katari → wide-interval expansiveness; little motion → syncopated busy rhythm; low register → high register; short chords → long chords.
• Workflow: make the sabi (C) and the introduction (A) first, and the B (middle section) last.

Endings of Setsu and Phrases (節や楽節の終わり方)

• How a phrase ends directly determines how the next phrase begins and the logic between them; how the setsu ends governs the whole song's aftertaste (後味).
• Closing (閉じて終わる) = full cadence on a stable tone: the statement is finished; the next section "starts anew." Opening (開いて終わる) = ending on an unstable tone: leaving resonance and lingering feeling (余韻); the next section "continues speaking."
• The eight-degree openness sequence (three axes: tonic → 3rd → leading tone × tonic chord → non-tonic chord → dominant chord × strong beat → weak beat via suspension/anticipation) → reference.md Table 5.
• Checklist of closing devices: append a closing phrase (終結フレーズ); repeat the final 2 bars (driving the point home, 念押し); repetition + augmented values (extremely strong cadence but easily shitsukoi/overbearing); deliberately append an incomplete cadential phrase for afterglow; neither melody nor harmony cadences (maximum afterglow); harmony-only cadence ("never quite finishing the sentence"); cut the cadence and charge straight into the next phrase (prevents dragging but sounds impatient).
• Kime (キメ, decisive tag): a flag-waving short phrase (~2 bars) appended to the sabi ending, often with the accompaniment stopping dead or a signature rhythm, crisply singing out the title/keyword; if built from a variant of the sabi's motive, unity is strong; a kime grown to phrase length = effectively a double chorus. When the song lacks meri-hari (メリハリ, dynamic contrast/edge), consider adding a kime.

Assembling the Whole Song

• Norm: 2–3 setsu, 3–5 minutes; slow songs reduce setsu count, fast songs increase it; the only criterion = listener satisfaction.
• Omit phrases in the latter half to defeat prediction: verse 2 no longer repeats A (A-A'-B-C → A'-B-C); verse 3 starts from the sabi run-up only (B-C).
• Repeat the sabi in the last setsu as the whole-song climax (A-B-C-C, even C-C-C-C); optionally with a modulation lifting the register.
• Atama-sabi (頭サビ, chorus-first opening): open directly with the sabi — verbatim = a splashy opening; thin accompaniment + slow tempo + rubato = a katari-style introduction.
• Pre-verse introduction: a recitation-like quiet opening (mid register + nearly static pitch + simple rhythm + barely moving harmony) or its opposite, a high-flying opening (leaps + high register).
• Ōsabi (大サビ, the added bridge between repeated setsu): a transitional section inserted between setsu repetitions — its character ≈ an introspective second B-melo, not a bigger climax; often feels half-time, low murmured asides; it must be followed by the sabi (after the calm, the sabi lands more indelibly). Not obligatory — an ōsabi added out of habit actually saps momentum.
• Coda (コーダ), four types: the grand-finale broad-melody type (= one more sabi at the song's end) / the pre-coda modulation type (receiving the whole song and sealing it) / the rubato·rit. calming introspective type / the A-melo recurrence type (slower tempo, thin accompaniment, repeating the opening material = a profound close by the recurrence principle).
• Intro (前奏) = the instrumental version of the introduction (usually borrowing A's or the sabi's harmony with a new melody; when the song already has a katari pre-verse or atama-sabi, the intro is often minimal or absent); deliberately no intro: a pickup melody entering on bare voice = raw and immediate (生々しい).
• Interlude (間奏) = the Second Introduction (verse 2's intro); interludes grow longer toward the song's latter half; often also an instrumental repetition of the setsu's final phrase.
• Outro (後奏) = the instrumental version of the coda; after it, another sung coda can still follow. Vamp (ヴァンプ) = a section presenting only the rhythm pattern, no melody.
• Designing the empty space: the melody need not carry everything — where the melody rests, listeners' attention turns to the accompaniment's rhythm and harmony; the build-up before the sabi can be delegated to the interlude/accompaniment; thinking about "the parts not sung" naturally clarifies what arrangement is needed.

11. Avoiding Cliché (予定調和)

• Yotei-chōwa (予定調和, predictable resolution / cliché predictability) = the listener reads the next move, everything unfolds as expected — the problem melody writing must guard against most.
• Checklist of high-risk patterns:
  • The same phrase repeated 3+ times without change;
  • A motive shifted in parallel along the chord progression (by the 3rd chord the listener has read the whole thing);
  • Formulaic harmonic rhythm: one chord per bar, dominant at the end of 8 bars;
  • Verse 2 copying verse 1's A−A' opening (the listener instantly predicts B−C must follow);
  • The long-short-long-short skip rhythm maintained throughout;
  • Textbook leading tone → tonic and iv → iii resolutions played straight.
• Countermeasures:
  1. The 3rd repetition must vary; in verse 2 omit the opening repetition (A-A'-B-C → A-B-C);
  2. Exploit expectation and turn the tables: first induce the listener's prediction, then at the predicted point betray it decisively (大きく裏切る) — staging the climax with surprise;
  3. Stretch the melodic line: a long arc not chopped up by motives (use anticipations to bind the rhythmic shapes, dive deep → the rebound ascent becomes the climax → a strong semitone appoggiatura as a second center of gravity → an unhurried close) — a long line = a sense of direction + a relay of varied motion devices, not mere absence of breaks (which would be dara-dara, dragging);
  4. Keep materials to a minimum: any new phrase must clarify "what differs from before, what continues" (part contrast + part repetition); songs with less material are more comprehensible and more moving;
  5. Break stock resolutions with suspensions/appoggiaturas/anticipations;
  6. Between "banal" (copying stock devices — reassuring but worthless) and "unintelligible" (straying so far from shared forms it becomes a wall), support originality (オリジナリティ) with strong logic internal to the work.
• Distortion + answer: all-smooth (スムーズ) = boring; you need an ibitsu (いびつ, warped/distorted) spot to stir the waters — but the distortion must be answered (a moving resolution after a run of unstable tones; an accent-defying leap = the endorsement of strong emotion), otherwise it is merely a defect.
• Earning trust: once earlier ingenuity convinces the listener "this song is worth listening to," even a flat phrase will be waited out patiently for "the good answer coming later."
• The ultimate standard: a 3-minute song consumes 3 minutes of the listener's life — every moment must be full, and finishing it should feel like having lived through a good experience.

12. The Creative Workflow (from Spark to Full Song)

Conceive the Whole First

• Section-by-section accretion ("waiting for the next good phrase to descend") jams up as constraints pile layer on layer: the second section is constrained by the first, the third doubly so… classic symptom = a gripping opening + a fizzled ending.
• The first fragment that surfaces is a good thing (it is the very motive force of composing), but soon afterwards answer: where does this song go? What development toward what climax? What is the theme? What about the introduction?
• The four items of conception:
  1. Musical: is there a use of tones "only this song has" (exceptionally fluid / exceptionally soft / one exquisite point of unease)?
  2. Content: what is sung, what drama (ドラマ) — the subject must be something everyone has experienced (empathizable), the telling must be one-of-a-kind (extraordinary); private imagery must be translated through public associations (a mayfly = evanescence works; an insect nobody knows = hitori-yogari 独り善がり, self-indulgence);
  3. Overall mood (全体のムード): ultimately cashed out in concrete elements — tempo, rhythm pattern, key and mode, degree of functional-harm

Truncated - read the full file at https://github.com/jtydhr88/japanese-composition-skills/blob/fe52df3f31625e911b654060130433f43b620b76/plugins/japanese-composition/skills/jp-melody-writing/SKILL.md.