Recognition Physics φ · PMNS Lab

Neutrino mixing, scored against NuFIT. Tension included.

Recognition Science writes the three neutrino mixing angles and the CP phase as fixed φ- and α-expressions, preregistered rather than fitted: sin²θ₁₃ = φ⁻⁸, sin²θ₁₂ = φ⁻² − 10α, sin²θ₂₃ = ½ + 6α, δCP = π + (6/11)(π/4) ≈ 204.5° (the scorecard object deltaCP_pmns_torsion_correction, with a proved range theorem; the same module also still carries predicted_deltaCP ≈ 191°). This script scores all four against NuFIT 5.2 with χ², in both θ₂₃ octants. Two corrections this page owes its own history: an earlier summary line contradicted its own table, and an earlier δCP value (11/8·π, never formalized in Lean) has been retired in favor of the scorecard prediction. The remaining tension is the content: θ₂₃ wants the upper octant, and DUNE and JUNO will settle it.

pmns_chi2.py · needs numpy + scipy · run verified 2026-07-08

−1.46σθ₁₃ = φ⁻⁸ (best)
p = 0.002angles, upper octant
+0.45σδCP = 204.5° (Lean)
numpy+scipy~1s runtime
Preregistered predictions

φ-harmonics, no fit

sin²θ₁₃ = φ⁻⁸ = 0.02129 sin²θ₁₂ = φ⁻² − 10α = 0.30899 sin²θ₂₃ = ½ + 6α = 0.54378 δCP = π + (6/11)(π/4) = 204.5°
What the data say today

Not falsified. Under real pressure.

θ₁₃ sits within 1.5σ: the strongest hit. θ₁₂ is comfortable at 0.5σ, and the Lean δCP lands 0.45σ from the NuFIT center (which still carries a large error). θ₂₃ is the pressure point: the prediction needs the upper octant, and the octant is experimentally unresolved (the two octant fits disagree at +5.6σ vs −3.5σ).

Kill conditions

Written into the script

sin²θ₁₃ outside [0.0190, 0.0235]; a definitive lower octant for θ₂₃; or DUNE finding δCP far from 204.5° at design precision, or anywhere outside (π, 2π) where the Lean range theorem puts it. Each one alone kills this formulation.

The pulls, drawn

Each bar is the deviation of one prediction from NuFIT 5.2 in standard deviations. The two θ₂₃ bars are the same prediction scored against the two octant fits; the experiment has not yet chosen between them.

z-scores vs NuFIT 5.2. Bands: |z| < 1 (dark), |z| < 3 (light). The full-test χ² = 34.21 in the reference output scores against the lower octant, where the θ₂₃ tension is largest; the upper-octant angles-only test gives χ² = 14.82 (p = 0.002). The retired 11/8·π value for δCP (never in Lean) sat at +2.36σ and is shown struck through.

χ² table (upper octant; δCP octant-independent)

ObservableRSNuFITzχ²
sin²θ₁₃0.021290.02225 ± 0.00066−1.462.13
sin²θ₁₂0.308990.303 ± 0.012+0.500.25
sin²θ₂₃0.543780.572 ± 0.008−3.5312.44
δCP/π = 25/22 (Lean)1.136361.08 ± 0.125+0.450.20
δCP/π = 11/8 (retired, not in Lean)1.3751.08 ± 0.125+2.36

Run it yourself

$ python3 pmns_chi2.py
Angles only: χ² = 14.82 for 3 dof, p = 0.0020
Full test:   χ² = 34.21 for 4 observables, p = 0.000

INTERPRETATION:
  • The mixing angles show tension (p = 0.002), dominated by the θ₂₃
    octant; the octant is experimentally unresolved and DUNE will decide it
  • The θ₁₃ = φ⁻⁸ prediction is the strongest (within ~1.5σ)
  • The Lean δCP prediction (25/22·π ≈ 204.5°) sits ~0.5σ from NuFIT

VERDICT:
  RS PMNS predictions are NOT YET FALSIFIED but face tension on the θ₂₃
  octant. Future precision experiments (DUNE, JUNO) will be decisive.
↓ PMNS χ² script Reference output
Honest scope

HYPOTHESIS with named falsifiers. The angle formulas (φ⁻⁸, φ⁻² − 10α, ½ + 6α) are CONJECTURE/MODEL in MixingDerivation; Lean then proves absolute-error match certificates to PDG centers (pmns_theta*_match), packaged as PARTIAL_THEOREM in PMNSScoreCard, not uniqueness of the formulas. The δCP value π + (6/11)(π/4) is the scorecard / SMCompletenessCert object (PMNSMatrix.deltaCP_pmns_torsion_correction, proved range π < δCP < 2π); the same module also still defines predicted_deltaCP = π + (φ−1)π/10 ≈ 191°, so the library is ambiguous at the definition layer and only the scorecard prefers torsion. The torsion derivation narrative is MODEL-tier. The α entering the angles is the RS construction value via the canonical exponential pipeline, which itself misses CODATA by 5.6 ppm (see the α lab); the exact fine-structure constant remains OPEN. A correction this page owes you: an earlier version tested δCP/π = 11/8, a script-level reading that was never formalized; it sat 2.4σ high and has been retired for the scorecard value, which sits at 0.45σ. Current data put real pressure on θ₂₃. This page exists to show the pressure, not to spin it. A definitive lower-octant θ₂₃ or a precise δCP away from 204.5° ends this formulation, and we will report that here if it happens.

All labs Regge triangulation Bekenstein spectrometer Lattice Lorentz Six gravity numbers 3/8 crossing Mass ratio PMNS χ² Galaxy rotation Regge cubic Solar eigenvalues Erdős–Straus The α construction