The Lattice-Prime Conjecture

An informal write-up of a conjecture about how small primes organize the exceptional structures of mathematics through a hierarchy of lattice spaces.

The core idea

The space of all 2-dimensional lattices (parametrized by the j-function on PSL(2,ℤ)\ℍ) generalizes to higher dimensions through Siegel modular forms and Shimura varieties. At each level, the moduli space has special points — elliptic points — corresponding to lattices with extra symmetry. The conjecture is that these special points are organized by the small primes in a very specific way, and that the exceptional structures of mathematics (Monster, sporadic groups, exceptional Lie algebras) are what you get when you crystallize the continuous lattice space at these special points.

The hierarchy

Each level corresponds to a normed division algebra dimension:

Level 1 (real, dim 1): The integers themselves. Galois theory of ℚ. Primes appear as the irreducible factorizations. The "Galois lattice" of all integer divisions of a circle.

Level 2 (complex, dim 2): 2D lattices, parametrized by τ in the upper half-plane mod PSL(2,ℤ). The j-function maps this to ℂ. Two elliptic points only: j(i) = 1728 (square lattice, ℤ[i]) and j(ω) = 0 (hexagonal lattice, ℤ[ω]). These correspond to the first two Heegner discriminants Δ = -4 and Δ = -3. The Monster lives here, with monstrous moonshine encoding the j-expansion. The supersingular primes (the 15 primes dividing |𝕄|) live here, terminating at 71.

Level 4 (quaternionic, dim 4): 4D symplectic lattices, parametrized by 2×2 symmetric matrices with positive-definite imaginary part. This is the Siegel upper half-space of genus 2. The moduli space 𝒜₂ has its own elliptic points. The most symmetric 4D lattice is D₄, whose vertex configuration is the 24-cell. Claim: The 24-cell honeycomb unifies the square and hexagonal cases at level 2 in the same way that ℤ[ω] and ℤ[i] are both subsumed by the quaternionic structure.

Level 8 (octonionic, dim 8): This level should host exceptional Lie group automorphic forms (E₈, etc.). The Leech lattice in 24D = 3×8 lives in extensions of this. This level is much less developed in the standard literature.

The small primes do the work

At each level, certain small primes show up as the irreducible obstructions that crystallize the continuous moduli space into discrete exceptional structure:

• 2, 3 — generate the 3-smooth world: tetrahedron, octahedron, cube, hexagonal tiling, 24-cell, F₄ root system.
• 5 — enters at the icosahedral level. Forces A₅ (smallest non-abelian simple group), the 600-cell, H₃ and H₄ Coxeter groups, the unsolvability of the quintic. Appears at level 4 as 5² 24-cells in the 600-cell.
• 7 — enters at the octonionic level. Fano plane, Klein quartic, G₂, Hurwitz surfaces, M₂₄.

The Heegner discriminants 1, 2, 3, 7, 11, 19, 43, 67, 163 correspond to the imaginary quadratic fields with unique factorization. Their differences are 1, 1, 4, 4, 8, 24, 24, 96 — exactly the normed division algebra dimensions (1, 4, 8, 24) with repetitions and a final factor of 4. The complex dimension (2) is skipped because complex symmetry is continuous, not discrete.

The kissing-number-plus-one observation

For each dimension n where the kissing number is known or conjectured, the kissing number plus one (adding a central sphere) is almost always prime:

The "exception" at dimension 4 (where 24 + 1 = 25 = 5²) is exactly where the next prime layer (5) enters the geometric story via the 600-cell.

The Brown number termination

The known solutions to n! + 1 = m² (Brown numbers) are:

• 4! + 1 = 5² (the 3-smooth → 5-smooth transition)
• 5! + 1 = 11² (5-smooth contains 11 as a higher symmetry)
• 7! + 1 = 71² (7-smooth contains 71)

71 is also:

• The largest supersingular prime
• The largest prime dividing |𝕄|
• One of the three primes whose product 47 × 59 × 71 = 196883 (smallest non-trivial Monster representation)

Conjecture: No further Brown numbers exist, and the reason is structural rather than probabilistic. The sequence 5, 11, 71 marks the saturation of "nice symmetry" at the levels indexed by 4!, 5!, 7!. Beyond the Monster, no further exceptional structure exists that could generate a fourth Brown number.

The (p² − 1) pattern

For any prime p ≥ 5, p² − 1 is divisible by 24. For small primes:

• 5² − 1 = 24 — the 24-cell, F₄
• 7² − 1 = 48 — the 288-cell (24-cell + dual), F₄ root system
• 11² − 1 = 120 — the 600-cell, S₅
• 13² − 1 = 168 — the Klein quartic, Fano plane, PSL(2,7)
• 17² − 1 = 288 — the 288-cell again
• 19² − 1 = 360 — A₆

These keep landing on dimensions or orders of exceptional structures. The reason (p² − 1)/24 is always an integer for primes ≥ 5 is elementary, but the fact that the specific values keep being structurally meaningful is the conjecture's content.

The fundamental claim

Each prime p represents a kind of irreducible obstruction in the closure of structure. Cyclic groups ℤ/n are the "discrete return distances" needed to close a loop, and prime ℤ/p are the irreducible such distances. As structures of higher complexity try to close on themselves, they generate residual twists that require larger primes to express.

The continuous lattice spaces (j-function, Siegel moduli, etc.) describe the smooth landscape. The exceptional structures are the discrete fixed points where this landscape crystallizes, indexed by the small primes that can serve as boundary integers in low-dimensional geometric enumerations.

The conjecture predicts:

1. The Monster is maximal because g = 1 (the elliptic-curve case) is the simplest non-trivial level, and 71 is the largest prime that can be a structural index there.
2. Higher Siegel/Shimura levels exist but don't produce a "larger Monster" — they produce more abstract exceptional structures (E₆, E₇, E₈ automorphic forms, exceptional groups over the octonions) but no further sporadic groups beyond what's already known.
3. The level past octonions (sedenions and beyond) is where the division algebra property fails, and exceptional structures cease to crystallize cleanly.

What's known vs. what's speculative

Established:

• The j-function, Siegel modular forms, Shimura varieties as the higher-dimensional generalization.
• The Monster's connection to the j-function (monstrous moonshine, proved by Borcherds).
• The supersingular primes coinciding with prime divisors of |𝕄| (proved by Ogg, generalized by Nakaya 2018 to the Baby Monster and Fischer's group via class numbers).
• p² − 1 divisible by 24 for p ≥ 5.
• The Heegner numbers being exactly 1, 2, 3, 7, 11, 19, 43, 67, 163.
• Brown numbers known to be only (4,5), (5,11), (7,71); Erdős conjectured no more; abc would imply finiteness.

Speculative / synthesis:

• The Heegner difference sequence 1, 1, 4, 4, 8, 24, 24, 96 encoding the normed division algebra dimensions with the complex dimension skipped.
• The Brown numbers terminating at 71 specifically because of Monster saturation.
• The kissing-number-plus-one pattern being a manifestation of the same crystallization mechanism.
• The 24-cell unifying hexagonal and square cases at the 4D level the way j(i) and j(ω) appear at the 2D level.
• The whole structure being one "higher-dimensional Galois theory of lattices" organized by the normed division algebra hierarchy.

Tests and consequences

If the conjecture is right, then:

• No fourth Brown number exists, and any future proof of this would route through the structural saturation of the Monster rather than through pure analytic number theory.
• The exceptional structures at higher Siegel levels (g = 2, 3, ...) should organize around small primes in patterns analogous to the j-function case, with the relevant primes shifting as the dimension changes.
• The deep connections between number theory (Heegner numbers, class field theory), geometry (lattice symmetries, Coxeter groups), and physics (string theory, holographic principle) are not coincidences but reflections of a single mechanism: discrete primes as crystallizations of continuous lattice moduli.