#4563 Knowledge Graph of Musical Chords

#4563 Knowledge Graph of Musical Chords

#4563 Describe a musical chord knowledge graph

A musical chord knowledge graph is a structured, multi-dimensional semantic network that maps the relationships between chords, their constituent notes, acoustic properties, and functional roles within harmonic theory.

Instead of treating chords as isolated text labels or static MIDI arrays, a knowledge graph models harmony as an interconnected web of nodes and edges, allowing a system to traverse musical concepts fluidly—such as moving from a specific pitch collection to its functional role in a jazz progression.

Here is a breakdown of how such a graph is architected, from its core ontology to its practical applications.

1. Core Ontological Nodes (Vertices)

To capture the full depth of harmonic theory, the graph requires a multi-layered node strategy:

• Pitch Class Nodes: The fundamental building blocks ($C, C\sharp/D\flat, D, \dots$). These are abstract entities independent of octave.

• Pitch Nodes (Notes): Specific frequencies mapped to MIDI note numbers ($C4, E4, G4$).

• Chord Prototype/Type Nodes: Abstract harmonic structures defined by intervals rather than root notes (e.g., Major Triad, Minor 7th, Dominant 7th ♯9).

• Chord Instance Nodes: A specific chord type bound to a root pitch class (e.g., C Maj7, A min7).

• Voicing/Inversion Nodes: The concrete spatial arrangement of notes across octaves (e.g., C Maj7 Drop 2, G7 1st Inversion).

• Scale/Mode Nodes: Collections of pitch classes that establish harmonic context (e.g., C Ionian, A Aeolian, G Mixolydian).

• Functional/Roman Numeral Nodes: The structural role of a chord relative to a key (e.g., I, ii, V7).

2. Relationships (Edges) and Properties

The real power of the graph lies in its typed, directed edges, which define the syntax of music theory:

Membership & Structure

• HAS_PITCH_CLASS: Connects a Chord Instance to its base pitch classes (e.g., C Maj7 --> C, E, G, B).

• HAS_INTERVAL: Connects a Chord Prototype to its internal interval formula (e.g., Minor Triad --> Root, Minor 3rd, Perfect 5th).

• IS_INVERSION_OF: Connects a specific voicing node back to its parent Chord Instance.

Harmonic Context & Function

• MEMBER_OF_SCALE: Links a chord instance to the scales it naturally occurs within (e.g., D min7 --> C Major Scale).

• RESOLVES_TO: Captures voice-leading tendencies and functional gravity (e.g., G7 --> C Maj, G7alt --> C min).

• SUBSTITUTE_FOR: Connects chords that share similar harmonic functions (e.g., D♭7 (Tritone Substitution) --> G7).

Voice Leading & Proximity

• VOICE_LEADING_DISTANCE: A weighted edge calculating the total semitone movement required to transition between two specific Voicing nodes (e.g., minimal smooth movement in voice-leading parsimony).

3. High-Density Graph Architecture (Schema Example)

If implemented in a property graph database like Neo4j (using Cypher) or an RDF triple store, a localized fragment of the schema looks like this:

(:Scale {name: "C Major"}) 
    -[:HAS_DIATONIC_CHORD {degree: "V"}]-> (:ChordInstance {name: "G7"})
    -[:IS_INSTANCE_OF]-> (:ChordPrototype {type: "Dominant 7th"})

(:ChordInstance {name: "G7"})
    -[:HAS_PITCH_CLASS]-> (:PitchClass {name: "G"})
    -[:RESOLVES_TO {type: "Authentic"}]-> (:ChordInstance {name: "C Major"})

(:Voicing {name: "G7 Drop 2"})
    -[:REPRESENTS]-> (:ChordInstance {name: "G7"})
    -[:LEADS_TO {semitone_cost: 2}]-> (:Voicing {name: "C Maj7 Root Position"})

4. Engineering Applications

Modeling chords in this manner unlocks complex algorithmic capabilities that traditional relational databases or flat files struggle to compute efficiently:

• Algorithmic Sound Design & Generative Composition: By querying nodes with a VOICE_LEADING_DISTANCE below a certain threshold, a Python audio engine or algorithmic sequencer can compute ultra-smooth, organic chord transitions on the fly.

• Context-Aware Chord Recommendation: An AI agent or local LLM can traverse the graph using path-finding algorithms (like Dijkstra's) to find unexpected harmonic paths (e.g., modulation via pivot chords or tritone substitutions) between two disparate keys.

• Acoustic Dissonance and Tension Analysis: By storing properties like consonance/dissonance ratings or intervallic vectors on the ChordPrototype nodes, an engine can mathematically calculate the tension curve of a progression by analyzing the properties of the traversed path.

• Interactive Musicology Wikis: Integrating this graph topology into a personal knowledge base (like Obsidian via semantic Markdown) allows for a dynamic "Living Notebook" where deep musical analysis is automatically cross-referenced by scale degrees, modes, and voice-leading rules.

From <https://gemini.google.com/app/>  Google Gemini (3.5 Flash)