Hyperbolic Flavor Geometry · Article I

The Smallest
Possible Universe

If space can curve negatively, how small can an entire universe become? And what does that minimum know about the mass of the muon?

Scroll to explore
0.9813688289
The volume of the smallest closed hyperbolic universe that can exist.
Not approximately. Not currently known. Proven.
The Problem

Why is the muon 207 times heavier than the electron?

The muon is a fat electron. Same charge. Same spin. Same forces. The only difference is its mass: 206.77 times heavier. The Standard Model describes the muon with extraordinary precision — predicting its magnetic moment to eleven decimal places — and offers absolutely no explanation for that factor of 207.

That number is inserted by hand. Measured, plugged in, never derived. It is one of approximately nineteen free parameters in the Standard Model that have no theoretical origin whatsoever.

207m_μ / m_e
3477m_τ / m_e
19Free parameters
"Why does matter come in three generations, each heavier than the last, with masses that seem arbitrary? This is the flavor problem — and it has no answer within the Standard Model."
Contrast — Hyperbolic Flavor Geometry

HFG reproduces the leptonic CP phase, the electron–muon mass ratio, the electron–tau mass ratio, and all three PMNS lepton mixing angles (θ₁₂, θ₂₃, θ₁₃) from the geometry of a single hyperbolic 3-manifold — with zero adjustable parameters.

What if the answer is geometric? What if the mass ratios, the mixing angles, the CP-violating phase are not arbitrary numbers but invariants of a specific mathematical space — fixed by geometry, not by accident?

This article is about the evidence that this is actually the case.

The Geometry

Curved space that closes on itself

In the nineteenth century, mathematicians proved that Euclid's geometry is not the only consistent geometry. There exists a geometry of constant negative curvature — hyperbolic geometry — where triangles have angle sums less than 180° and space expands exponentially as you move outward.

A compact hyperbolic 3-manifold is a finite chunk of this curved space, closed on itself like a doughnut but with hyperbolic geometry inside. These objects obey a theorem that has no flat-space analogue:

If two complete, finite-volume hyperbolic 3-manifolds are topologically equivalent, they are geometrically identical. Every geometric invariant — volume, geodesic lengths, curvature data — is completely determined by topology. There are no free parameters.

This is why hyperbolic 3-manifolds can potentially encode physical information without any tuning. The encoding is exact, or it is not an encoding at all.

Interactive — Poincaré Disk Model Move mouse · watch geodesics
The entire hyperbolic plane compressed into this disk. Straight lines are circular arcs meeting the boundary at right angles. Equal-sized shapes near the edge look smaller — but they are the same size in hyperbolic geometry. The boundary is infinitely far away.

Euclidean Space

Parallel lines stay parallel. Area grows as r². Space feels familiar.

Hyperbolic Space

Parallel lines diverge. Area grows as cosh(r)−1 ≈ e^r/2. Space expands explosively.

Interactive Demo

How fast does each space grow?

Drag the slider to increase the radius. Watch how Euclidean and hyperbolic areas diverge. This single demonstration explains why hyperbolic manifolds are so rigid and information-rich.

Radius vs Area Comparison

r = 0.00
0.00
Euclidean area πr²
0.00
Hyperbolic area 2π(cosh r − 1)
The Connection

The manifold that fits the data

At the bottom of the ordered census of compact hyperbolic 3-manifolds — the manifold with the smallest possible volume — sits the Meyerhoff manifold, volume 0.9813688289. It is constructed from m003 by Dehn surgery along the slope (−2,3).

A systematic scan of the SnapPy census, optimizing a fitness function measuring how well the holonomy group reproduces the observed PMNS lepton mixing matrix, found the global minimum at this manifold. Fitness: 0.005087. Free parameters: zero.

δ_HFG = 195.91°  |  δ_PDG = 197.0° CP phase · 0.55% discrepancy · zero free parameters

The lepton mass ratios appear as Eisenstein norms — exact algebraic integers in the number field generated by the manifold's cusp shape:

N(16 + 12ω) = 208 ≈ m_μ/m_e    error 0.59% ω = e^(2πi/3) · Eisenstein integer · same field as the cusp shape of m003
N(68 + 37ω) = 3477 = m_τ/m_e Four significant figures · error 0.006% · zero free parameters
M_PMNS = m003(−2,3) = m019(2,1)
vol = 0.9813688289 = v₀
H₁ = ℤ/5  |  fitness = 0.005087
Gal(cusp) = ℤ/2 = Weyl(SU(2))
δ_CP = 195.91° (PDG: 197.0°)
M_CKM = m006(−5,2)
vol = 2.0289  |  H₁ = ℤ/5
fitness = 0.016482 (CKM matrix)
Gal(cusp) = S₃ = Weyl(SU(3))
zero free parameters
The Arithmetic Core

Three shapes. One orbit. One invariant.

In the number field K = ℚ(w), w⁴ = w+1, the tetrahedral shapes of m019 are explicit units. They orbit under T(z) = 1/(1−z). The Bloch-Wigner value never changes.

z_A = w³ = 1−u₁    z_B = −w = u₁⁻¹    z_S = w⁻⁴ = u₁⁴

T(w³) = −w, T(−w) = w⁻⁴, T(w⁻⁴) = w³  (pure algebra from w⁴=w+1)
Product: w³ · (−w) · w⁻⁴ = −1 exactly.
D(z_A) = D(z_B) = D(z_S) = v₀  (Bloch-Wigner functional equation)
The Volume Spectrum

The discriminant −283 family

Every cusped hyperbolic 3-manifold in the SnapPy census with cusp field discriminant −283 has volume that is an exact rational multiple of v₀. All six known examples. All verified to machine precision.

Click any bar to see details.

vol(M)/v₀ ∈ ℚ for every disc=−283 cusped manifold in the census.
Sequence: 3, 4, 4, 5, 11/2, 6. Errors ≤ 8.88×10⁻¹⁶ (machine epsilon).
The Question

What is the geometry trying to tell us?

The minimum-volume hyperbolic universe encodes the PMNS mixing matrix. Its cusp-shape Galois group is the Weyl group of SU(2). Its tetrahedral shapes are units in the discriminant −283 number field, forming a period-3 orbit that forces their Bloch-Wigner values to equal its own volume. Every manifold in its arithmetic family has a volume that is a rational multiple of that volume.

The tau-electron mass ratio is 3477. The Eisenstein norm N(68 + 37ω) is 3477. These are not approximations — they are exact mathematical facts that are asking to be understood.

Whether this is a deep connection between arithmetic geometry and particle physics, or an elaborate numerical coincidence, remains to be determined. The mathematics is real. The programme continues.

6/6Census manifolds with rational vol/v₀
2×10⁻⁵¹Precision of D(w³) = v₀
0Free parameters in all predictions

Marvin L. Gentry · Independent Researcher · Seattle WA
Preprints: SSRN · Code: GitHub · Website: hyperbolicflavorgeometry.org