Extensions

The parameter n

The general structural development uses G_n = ℤ/nℤ. Setting n = 12 gives the chromatic universe; the 223 TI-classes, the Forte numbers, and the familiar chords and scales are consequences of that particular choice. But the structural theorems are n-agnostic:

• The finite Wiener–Khinchin correspondence holds for any n.
• A shift period of a gap word is equivalent to Fourier support in its specified annihilator subgroup. Arbitrary sparsity alone does not establish a proper period; this general theorem is not yet in Lean.
• The layered symplectic form ω_n,k is exact on ℂⁿ⁻¹ × ℂ^k for every (n, k).
• Transposition and cyclic reindexing have Hamiltonian circle extensions; exact root-forgetting is the finite C_k quotient. Inversion remains anti-symplectic.
• Cardinality indexes a disjoint family of spaces; no attachment topology across cardinalities is specified.

What does change with n is the combinatorial landscape: the number of TI-classes, the density of the quotient graph, the fiber sizes, and the symmetry-order distribution all depend on the arithmetic of n.

Other pitch systems

Choosing a different equal division changes the finite cyclic group, its class counts and its possible orderings. The structural arguments can be restated with their hypotheses, while each concrete census requires its own computation or proof. The released exhaustive census is twelve-tone.

A different repeating interval can also provide a cyclic model: an equal division of a tritave can be modeled modulo that tritave. Unequal tunings and pitch systems without such a repeating identification need separate choices of space, equivalence and distance. They are not all instances of the current finite cyclic model.

Just-intonation ratios and continuous pitch raise additional questions about group structure, topology and analysis. No general infinite-dimensional symplectic extension is established by the current discrete Lean release. Preserving timing, articulation and register for an instrument is an immediate practical problem even before these wider mathematical extensions.

Rhythm, time, and the second circle

Music has two fundamental circles: pitch (the octave) and time (the metric cycle). The same layered content/order architecture could potentially apply to rhythmic patterns on ℤ_n, where n now indexes time points modulo a metric period rather than pitch classes modulo the octave.

This is not merely speculative. Yust's Organized Time (already cited in GAMUT's bibliography) applies Fourier analysis to rhythm and meter, treating rhythmic patterns as subsets of a time circle and analyzing their spectral content. The layered extension would add an order layer — rooted cyclic orderings of onsets, gap words measuring inter-onset intervals, and a fiber bundle over rhythmic content classes — and then ask which additional geometric structure, if any, serves the intended musical task. Rhythmic analogues should distinguish onset-content autocorrelation, gap-word periodicity, and faithful order reconstruction, just as the pitch-class model does. This direction is unexplored but structurally natural.

A still more ambitious extension would consider pitch-rhythm interaction: a product of two cyclic groups, ℤ_m × ℤ_n, encoding simultaneous pitch and onset information. A model would need to preserve the association between each pitch and its onset; independent pitch and rhythm summaries can lose that pairing. A product construction alone does not settle the representation or its musical usefulness.

Connections to other fields

Crystallography and phase retrieval

Homometric sets have identical autocorrelation. Z-related pairs additionally exclude transposition/inversion equivalence. In crystallography, the same phenomenon appears as the phase problem in X-ray diffraction: the diffraction pattern records only the squared magnitudes of the Fourier transform of the electron density, losing phase information. Two crystal structures can produce identical diffraction patterns yet have different atomic arrangements. The mathematical core is the same: autocorrelation determines the power spectrum but not, in general, the underlying structure.

Signal processing and engineering

The Wiener–Khinchin theorem and cyclic autocorrelation are standard tools in signal processing, communications engineering, and radar. Duncan's original paper appeared in the Journal of the Audio Engineering Society, not a music-theory journal — reflecting the fact that the mathematical content lives at the intersection of discrete harmonic analysis and applied engineering. GAMUT's symplectic layer adds a geometric interpretation that is proposed for this layered model; its individual ingredients (action-angle variables, moment maps, Hamiltonian symmetries) are well-established in physics and engineering.

Combinatorics and group theory

The fiber construction is a special case of studying orbits and stabilizers under group actions on finite sets. The content layer is the orbit space of the dihedral group acting on subsets; the order layer is the orbit space of the cyclic group acting on permutations of a fixed subset. The Burnside counts, the stabilizer orders, and the stratification by cardinality are all instances of standard combinatorial group theory. What is new is the particular combination — content orbits, rooted cyclic order fibers, Fourier lifts of both — and the observation that this combination lands in symplectic territory.

Open problems and future directions

• Registral information and voice leading. The current framework is octave-equivalent: it does not distinguish between a C4 and a C5. Adding registral data would connect GAMUT to the orbifold chord spaces of Tymoczko and Callender–Quinn–Tymoczko, where voice-leading distance is the primary geometric structure. The challenge is to integrate this with the symplectic layer without losing the clean action-angle decomposition.
• Musically motivated Hamiltonians. The kinematics — the symplectic form, the symmetry actions and conditional reduction — are derived in the reviewed manuscript, not verified in Lean. But the framework does not yet contain a dynamics: there is no Hamiltonian whose flow models compositional preference, voice-leading cost, or harmonic tension. Defining such Hamiltonians, calibrating them against real repertoire, and studying the resulting flows is one open mathematical direction. The immediate instrument agenda instead starts with a small performer task and an audible baseline.
• Nonlinear spectral embeddings. The current content renderings use multidimensional scaling or linear projections. Diffusion geometry (Coifman–Lafon) and other nonlinear spectral methods could produce lower-dimensional embeddings that better preserve the graph topology of the content quotient.
• Computational scaling. Exact fiber enumeration is feasible through cardinality 5 (24 orderings per fiber); by cardinality 12 the fiber has nearly 40 million elements. The current implementation samples at most 96 per fiber. More sophisticated sampling, approximation, or algebraic strategies could extend the computational reach.
• Empirical validation. The geometric framework makes structural predictions (e.g., about the topology of fiber neighborhoods, the distribution of entropy across orderings, the spectral signatures of stylistically coherent pattern families). Testing these predictions against corpora of real compositions would ground the theory in musicological evidence.