The Standard Model contains exactly three generations of fermions: (e, nuₑ, u, d), (mu, nuₘu, c, s), and (tau, nuₜau, t, b). This replication of the electroweak multiplet structure three times over, with enormously different masses but otherwise identical quantum numbers, is one of the Standard Model’s most conspicuous unexplained features — what Rabi famously called ‘who ordered that? ’ on learning of the muon. Precision measurements of the Z-boson invisible width confirm exactly three light neutrino species (LEP 2006), and LHC searches have excluded standard fourth-generation quarks to high mass scales. The Standard Model accommodates this empirical fact but offers no reason for the number three rather than two, four, or more. We show that in the Three Time Dimensions (3+3) spacetime framework (de Haan 2026, book manuscript, DOI 10. 5281/zenodo. 19633127), in which the third time dimension t₃ is compactified as a discrete two-sphere S² with 2¹52 Planck-area cells, the number of fermion generations is forced to be exactly three by the same topological chain that forces the number of colour charges to be exactly three. Both follow from the Euler–Poincaré constraint: any hex-pentagon tiling of S² has exactly 12 pentagonal defects; these 12 defects admit a unique trisection with Nc (Nc+1) = 12; the unique positive-integer solution is Nc = 3. Each of the three trisection sectors supports one set of fermionic excitation modes — one generation. A fourth generation would require 20 defects (Nc = 4), which is topologically forbidden. The paper has two distinct results. The first, the central elegance claim, is the count: exactly three generations as a topological theorem, not an empirical input. The second is the mass hierarchy: the electron is 200x lighter than the muon which is 17x lighter than the tau because the three trisection sectors occupy different angular positions on S², and the Yukawa coupling varies as |P₂ (cos theta) | from zero at the nodal cone thetaₙode = arccos (1/sqrt (3) ) ~ 54. 74 degrees to unity at the poles. The wild disparity in fermion masses reduces to angular distances on a sphere. The paper makes one firm empirical commitment: no fourth fermion generation will be found at any scale. LHC has excluded fourth-generation quarks to ~1 TeV; HL-LHC will extend this to ~2–3 TeV in the 2030s; FCC-hh will probe to ~30 TeV in the 2040s. Any positive observation of a fourth generation at any mass falsifies the topological derivation. The related claim, that the same Z₃ symmetry resolves several other Standard Model puzzles simultaneously (Nc = 3 colours, three generations, deltaPMNS = 194. 48 degrees, thetaQCD = 0), is developed in the companion Strong CP preprint.
C. R. (René) de Haan (2026) studied this question.