V2.621 - Generation Selection from Lambda
V2.621: Generation Selection from Lambda
Motivation
The number of fermion generations (N_gen = 3) is one of the deepest unexplained facts in particle physics. The Standard Model works for any N_gen; nothing in gauge theory, anomaly cancellation, or Yukawa structure requires exactly 3.
This framework predicts Ω_Λ = R(N_gen) from the field content. Since R depends on N_gen, we can invert: what value of N_gen does the observed Ω_Λ require?
Key Result
R(N_gen) is strictly monotonically decreasing (proven analytically: dR/dN has numerator -3062 < 0). Therefore at most ONE integer N_gen can match Ω_Λ_obs.
| N_gen | N_Weyl | N_eff | R | Tension with Planck |
|---|---|---|---|---|
| 1 | 15 | 68 | 1.103 | 57σ |
| 2 | 30 | 98 | 0.832 | 20σ |
| 3 | 45 | 128 | 0.688 | 0.4σ |
| 4 | 60 | 158 | 0.598 | 12σ |
| 5 | 75 | 188 | 0.537 | 20σ |
N_gen = 3 is the unique solution. N_gen = 2 is excluded at 20σ, N_gen = 4 at 12σ.
Exact Analytic Formula
δ(N) = -29/3 - 11N/12 = -(116 + 11N)/12
N_eff(N) = 38 + 30N
R(N) = (116 + 11N)√π / (3(38 + 30N))Solving R(N) = Ω_Λ_obs gives N = 3.028 — the observed cosmological constant picks out 3 generations to within 3% of an integer.
Bayesian posterior: P(N=3 | Ω_Λ_obs) ≈ 1.000 for uniform prior over N = 1…6.
Gauge Group Selection
With 3 generations fixed, scanning gauge groups:
| Theory | N_vec | R | Tension |
|---|---|---|---|
| SM: SU(3)×SU(2)×U(1) | 12 | 0.688 | 0.4σ |
| SU(5) broken → SM | 12 | 0.688 | 0.4σ |
| SO(10) broken (+ν_R) | 12 | 0.667 | 2.5σ |
| SM + extra U(1) | 13 | 0.715 | 4.1σ |
| SU(3)×SU(3)×U(1) (331) | 17 | 0.815 | 18σ |
| Pati-Salam | 21 | 0.877 | 26σ |
Only the SM gauge group (or GUTs that break TO the SM) survives. An extra Z’ is excluded at 4.1σ. Note: SU(5) GUT broken to SM gives the same IR spectrum and is indistinguishable.
Higgs Sector Selection
| Higgs model | N_scalars | R | Tension |
|---|---|---|---|
| 1 doublet (SM) | 4 | 0.688 | 0.4σ |
| 1 doublet + singlet | 6 | 0.678 | 0.9σ |
| 2 doublets (2HDM) | 8 | 0.669 | 2.1σ |
| 3 doublets | 12 | 0.652 | 4.5σ |
The SM with 1 Higgs doublet is preferred. The 2HDM (as in MSSM) is at 2.1σ — marginally disfavored.
Joint Scan
Scanning 42 (N_gen, gauge group) combinations: only 3 are viable within 2σ.
All three have the same IR spectrum: SU(3)×SU(2)×U(1) with 3 generations. The one exception is a Pati-Salam model with 5 generations (R = 0.681, 0.5σ) — but this requires 5 generations and 21 gauge bosons, which is excluded by collider data independently.
Neutrino Nature
| Type | Weyl/gen | Best N_gen | R | Tension |
|---|---|---|---|---|
| Majorana | 15 | 3 | 0.688 | 0.4σ |
| Dirac | 16 | 3 | 0.667 | 2.5σ |
Both select N_gen = 3, but Majorana neutrinos are preferred by 2.1σ over Dirac.
What This Means
The framework doesn’t just predict Ω_Λ — it answers why 3 generations:
- R(N) is monotone decreasing, so at most one integer works
- The continuous solution is N = 3.028 — nature chose the nearest integer
- Adjacent values (N=2, N=4) are excluded at 20σ and 12σ respectively
- The gauge group and Higgs sector are also (approximately) selected
This is a retrodiction, not a prediction — we already know N_gen = 3. But it’s a non-trivial consistency check: there was no guarantee that the observed Ω_Λ would land on ANY integer. The probability of landing within 0.4σ of an integer in the range [1, 6] is ~30%, so this is not extremely improbable but is consistent.
Honest Assessment
Strengths:
- N_gen = 3 uniquely selected with overwhelming statistical significance
- The analytic formula R(N) = (116+11N)√π / (3(38+30N)) is exact — no fitting
- Monotonicity guarantees uniqueness: at most one integer solution exists
- SM gauge group + 1 Higgs doublet preferred over all alternatives
Weaknesses:
- This is a retrodiction (N_gen = 3 is already known), so it’s a consistency check, not a prediction
- The “proof” assumes the framework is correct — it doesn’t independently derive N_gen = 3
- Pati-Salam with 5 generations also fits (excluded only by collider data, not by Ω_Λ alone)
- Higgs sector selection is weaker: singlet extension at only 0.9σ
- The framework counts only IR degrees of freedom; UV completion could change the story
The testable prediction: If a new light field is discovered (e.g., sterile neutrino, dark photon, axion), R will shift by a calculable amount. The framework predicts the EXACT Ω_Λ change per new species — this IS testable.
Files
src/generation_selection.py: Core computation (generation scan, gauge scan, Higgs scan, monotonicity proof)tests/test_generation_selection.py: 9 tests, all passingrun_experiment.py: Full 9-part analysisresults.json: Machine-readable output