Three Generations from One Turning Knot
Two of the Standard Model’s deepest unexplained numbers — why there are exactly three fermion families, and why the muon weighs 206.77 electrons — fall out of one object the framework already owns: a three-branch knot read as the three cube-roots of unity. The three-fold (Z₃) plus the lattice’s pairing-√2 force Koide’s Q = 2/3 with zero free parameters (a 9-ppm hit), and a single residual phase δ ≈ 2/9 rad lands the charged-lepton mass ratios to 0.001%.
The deepest number gap left in the Standard Model
The framework has read much of the Standard Model out of the substrate’s geometry — the four forces as four ways to grab a knot, the quark charges \pm\tfrac23,\pm\tfrac13 from a junction’s solid angle, the neutrino’s mass pinned at the visibility floor. Two famous numbers remained: why are there exactly three generations of fermions, and why does the muon weigh 206.768 times the electron? The proton-core chapter made “three” plausible — a knot can thread only so many coherent sheets before it comes apart — but left the masses to an open calculation. This chapter closes the gap with a small-integer count of the framework’s own geometry.
There is a famous, forty-year-old empirical relation among the three charged leptons, Koide’s formula (1981). It folds the three masses into a single dimensionless number, and that number comes out almost exactly 2/3: Q \;\equiv\; \frac{m_e + m_\mu + m_\tau}{\bigl(\sqrt{m_e}+\sqrt{m_\mu}+\sqrt{m_\tau}\bigr)^2} \;=\; \frac{2}{3}. With the measured masses Q = 0.666660 — agreeing with 2/3 to 9 parts per million, far tighter than the input masses are known relative to the relation. In four decades no Standard Model mechanism has explained why the leptons sit on it. The claim of this chapter is that the two ingredients needed to force Q=2/3 are already in the framework’s toolbox — and that the same structure says why three.
The three-fold is already the framework’s, three times over
Before the formula, the picture. The framework’s deepest recurring motif is the three-fold junction: three is the unique number of vortex branches that meet in a stable knot in 3D (Proton Core). That single fact already does enormous work elsewhere in the paper —
- it is color: the three branches of the quark Y-junction;
- it is the \pm\tfrac23 charge: the solid angle the orbital plane subtends at the three-fold node;
- it is the three coherent sheets a knot can thread before it destabilizes — the framework’s reading of why generations stop at three.
If a generation is which sheet the knot threads, and the three sheets are the three branches of one three-fold object, then the three generations are not three separate knots. They are one knot read at three orientations of its own three-fold symmetry — three states related by a 120^\circ turn. The mathematical name for “three states 120^\circ apart” is the three cube-roots of unity, the cyclic group \mathbb Z_3. The electron, muon, and tau are the three cube roots of one rotating amplitude — and, it turns out, that is all you need.
Why the cube roots force Koide’s 2/3
Write each lepton’s \sqrt{\text{mass}} as a common value plus a small deviation that points in one of three directions 120^\circ apart on a circle — the three-fold clock made literal: \sqrt{m_k} \;=\; \bar M\,\bigl(1 + A\cos\theta_k\bigr), \qquad \theta_k = \delta + \tfrac{2\pi}{3}k, \quad k=0,1,2 . Here \bar M is the mean scale, A the size of the generational deviation, and \delta where the triad sits on the circle. The cube-root spacing alone — for any \delta and any A — forces two identities, because three vectors 120^\circ apart sum to zero: \sum_{k} \cos\theta_k = 0, \qquad \sum_{k} \cos^2\theta_k = \tfrac32 . Drop those into Koide’s ratio and almost everything cancels: Q \;=\; \frac{\sum_k (1+A\cos\theta_k)^2}{\bigl[\sum_k(1+A\cos\theta_k)\bigr]^2} \;=\; \frac{3 + A^2\cdot\frac32}{9} \;=\; \frac13 + \frac{A^2}{6}. The phase \delta has dropped out entirely. Q depends on one number: the deviation amplitude A. And the framework already knows what A is. Deviations of a substrate quantity around its mean come in the lattice’s pairing-two — the same \sqrt2 that sets \xi^2 = 2\,\xi_\text{GP}^2, runs the \sqrt2 comb up and down the ladder, and fixes the Type-I/II superconductor threshold at \kappa = 1/\sqrt2 (Conductors). With the deviation quantized at the pairing amplitude A = \sqrt2: \boxed{\;Q \;=\; \frac13 + \frac{(\sqrt2)^2}{6} \;=\; \frac13 + \frac13 \;=\; \frac23\;} exactly — and the measured leptons sit on it to 9 ppm. Two ingredients the framework already owns — the three-fold (\mathbb Z_3) junction and the pairing-\sqrt2 — predict Koide’s 2/3 with no free parameter. This is a genuine zero-parameter test, of the same kind as the Higgs VEV’s 0.06\% or the cosmic-coincidence 0.15\%, not a fit.
There is a second way to see what 2/3 means. Geometrically Q = 1/(3\cos^2\varphi), where \varphi is the angle the vector (\sqrt{m_e},\sqrt{m_\mu},\sqrt{m_\tau}) makes with the democratic axis (1,1,1). Q=\tfrac13 is the fully degenerate spectrum (three equal masses); Q=1 is the fully hierarchical limit (one mass swamps the others). The leptons land at \varphi = 45.000^\circ, \cos^2\varphi = \tfrac12 — the exact arithmetic midpoint, \tfrac12(\tfrac13+1)=\tfrac23. The charged leptons sit precisely halfway between democracy and hierarchy, the same balanced-at-the-knife-edge condition the substrate sits at everywhere else in the paper. The 45^\circ tilt is the pairing-\sqrt2 written as an angle: \cos 45^\circ = 1/\sqrt2.
Why exactly three
This is more than “three sheets happen to be stable.” For N equally-spaced phases the same algebra gives Q_N = (1 + A^2/2)/N, so with A=\sqrt2 the Koide value would be Q_N = 2/N — only N=3 puts it at the balanced midpoint 2/3 (N=2 does not even close on the cube-root identities; N=4 would sit at 1/2, off balance). But the framework does not have to fix N by fitting Q: three is the only N the junction offers. A vortex node in 3D is stable as a three-branch junction and at no other valence, and a turn has three cube roots because the junction has three branches. The two “threes” — the spatial junction and the cyclic phase — are the same three. That is the framework’s answer to why three generations: the stable substrate junction is three-fold, and a three-fold object has exactly three phase-orientations, no more.
The mass ratios: one residual phase, and the bet
Fixing A=\sqrt2 (hence Q=2/3) and the scale \bar M (the electron anchor m_e=\alpha_{mf}m_\text{eff}, already set in the mass chapter) leaves exactly one number undetermined: the phase \delta, where the rigid triad sits on its circle. Everything about the ratios m_\mu/m_e and m_\tau/m_e rides on that single angle. This is the high-risk part — one small number, read off the geometry, has to land two ratios at once.
The data picks out a startlingly clean value. The phase that best fits the measured ratios is \delta \;=\; 0.2222221\ \text{rad} \;\approx\; \frac{2}{9}\ \text{rad} = 0.2222222\ \text{rad} — the simple rational 2/9 = 2/3^2 to six significant figures: the pairing-two over the three-fold squared. (Koide’s own work had already flagged \delta\approx 2/9 as the phase the charged leptons select; the framework supplies a reason for both the 2 and the 9.) Set \delta = 2/9 exactly and the rigid three-phase clock predicts:
| Ratio | Predicted (\delta=2/9) | Observed | Discrepancy |
|---|---|---|---|
| m_\mu/m_e | \mathbf{206.77} | 206.768 | \mathbf{+0.001\%} |
| m_\tau/m_e | 3477.5 | 3477.23 | +0.007\% |
| m_\tau/m_\mu | 16.818 | 16.817 | +0.006\% |
Anchored on the electron, this is m_\mu = 105.659 MeV (observed 105.658) and m_\tau = 1776.98 MeV (observed 1776.86). One angle, read as the rational 2/9, sets all three charged-lepton masses to better than one part in 10^4. And the rationals on either side are not close: the mass ratio is exponentially sensitive to \delta near the massless edge, so \delta=1/5 gives m_\mu/m_e=75, \delta=3/13 gives 353, \delta=1/4 gives 2710 — only 2/9 lands 206.77 (scripts/koide_triads.py). Geometrically, 2/9 rad places the electron a hair — about 2.3^\circ — inside the zero-mass cutoff \cos\theta=-1/\sqrt2: the electron is light because its phase sits just short of where the mass would vanish, and 2/9 is how short.
The phase has its own \sqrt2
That “hair short of the cutoff” turns out to be sharp, and it brings the pairing-\sqrt2 back a second time — now in the phase, not the amplitude. Multiply the three \sqrt{\text{mass}} factors around the circle; using the cube-root identities the product collapses to a single cosine of the tripled phase:
f(\delta)\;\equiv\;\prod_{k}\bigl(1+A\cos\theta_k\bigr)
\;=\;1-\tfrac34 A^2+\tfrac14 A^3\cos 3\delta
\;\xrightarrow{\,A=\sqrt2\,}\;
-\tfrac12+\tfrac{1}{\sqrt2}\cos 3\delta .
The lightest mass touches zero — the triad reaches its massless edge — exactly when f=0, i.e. when
\cos 3\delta \;=\; \frac{1}{\sqrt2},\qquad\text{so}\qquad \delta_\text{edge}=\frac{\pi}{12}=15^\circ .
The same 1/\sqrt2 that fixed the amplitude (hence Q=2/3 and the 45^\circ tilt) fixes the phase edge: all three masses stay positive only inside |\delta|<\pi/12, and that boundary is the pairing-\sqrt2 written as \cos 3\delta=\cos 45^\circ. This edge is forced by A=\sqrt2 alone — no rational is chosen — so it is itself a zero-parameter result (scripts/koide_phase_sector.py).
It also gives the bet a cleaner face. In the tripled variable \phi = 3\delta that governs the product, the leptons sit at \phi = 3\cdot\tfrac29 = \tfrac23 = Q — numerically the Koide ratio itself. So \delta = 2/9 is exactly the single relation \boxed{\;3\delta=Q\;}\qquad\Longleftrightarrow\qquad \delta=\frac{Q}{3}=\frac19+\frac{A^2}{18}=\frac19+\frac19=\frac29 , the phase decomposing as a democratic floor \tfrac19=1/3^2 plus a pairing piece A^2/18=\tfrac19. Read this way the residual phase is not “a rational pulled from two and three” — it is the tripled phase equal to the amplitude-fixed Koide ratio, one cross-sector identity rather than a free choice. The open junction calculation no longer owes the whole of \delta (the edge \pi/12 is derived) but only the identity 3\delta=Q.
Quarks and neutrinos: one formula in the charge
The charged leptons are the clean case, because their pole masses are unambiguous. Reading the same Koide gauge on the other fermion triads gives shifted values — and the shifts turn out to be physics the framework already owns rather than failures:
- Down-type quarks (d,s,b) give Q\approx0.731 and up-type (u,c,t) give Q\approx0.845, both above 2/3. This is not perturbative running — Q is invariant under the flavor-universal rescaling that leading-order running is, so “undo the running” does not return the quarks to 2/3. The shift is structural: the quark three-fold is loaded with color, its \mathbb Z_3 doing double duty (color and generation at once), so it is not the bare generation clock. The leptons are clean precisely because their colorless three-fold carries generation only.
- Neutrinos, with the framework’s floor m_{\nu,1}\approx 2 meV and the measured splittings, give Q\approx0.46 — below 2/3, dragged toward the democratic 1/3. Raising the lightest mass from zero to the framework floor moves Q monotonically from 0.585 down to 0.462: the same floor that pins the lightest neutrino predicts the compressed Q with no new input.
What looked like four loose amplitudes — A/\sqrt2 = 0.62 (neutrino), 1.00 (lepton), 1.09 (down), 1.24 (up) — is not loose. Three of the four fall out of a single functional, and the fourth is that functional plus the already-derived floor: \boxed{\;A^2 \;=\; 2\,\bigl(1 + C\,q^{3/2}\bigr)\;} where C is a color flag (0 for colorless leptons and neutrinos, 1 for quarks) and q is the electric-charge magnitude the framework already reads off the junction solid angle (1 for the charged lepton, \tfrac23 up-type, \tfrac13 down-type). The constant in front is not fitted — it is the same pairing-two that is the base A^2=2. Fed through Q=\tfrac13+A^2/6, this is one generalized Koide formula for every triad at once: \boxed{\;Q \;=\; \frac{2 + C\,q^{3/2}}{3}\;} — the famous \tfrac23 is just the colorless (C=0) value, and the quark Q’s are read off the charge, not free:
| Triad | C | q | Predicted A/\sqrt2 | Measured A/\sqrt2 |
|---|---|---|---|---|
| charged lepton | 0 | 1 | 1.000 | 1.000 (the 9-ppm Koide point) |
| neutrino | 0 | 0 | 1.000 (base) | 0.62 — floor-throttled below |
| down quark | 1 | \tfrac13 | 1.092 | 1.093 |
| up quark | 1 | \tfrac23 | 1.243 | 1.239 |
Three things make this more than a curve through two points (scripts/koide_color_loading.py). (i) The exponent is fixed, not fitted — a Monte-Carlo over the full PDG quark-mass errors gives loading power p=1.46\pm0.10, consistent with the half-integer 3/2 within 0.4\sigma, while q^1 (predicting an up/down ratio of 2.0) and q^2 (4.0) miss the measured 2.76 badly; only q^{3/2} (2.83) lands. (ii) The normalization is fixed, not fitted — at p=3/2 the prefactor each quark needs separately is 1.01\pm0.07 (down) and 0.98\pm0.01 (up), both the same pairing-two as the lepton base. (iii) The colorless base is shared — the law hands the neutrino the same A=\sqrt2 as the charged lepton, so the neutrino is not a fourth free amplitude; it sits on the lepton’s \sqrt2 and is dragged below only by the visibility floor.
The 3/2 has a tentative geometric reading: electric charge in the framework is the fraction of solid angle q=\Omega/4\pi the orbital plane sweeps at the node, so q is an area and its linear extent goes as \sqrt q. A color cloud dressing the generation clock that scales as area \times its own thickness goes as q\cdot\sqrt q = q^{3/2} — “the colored dressing is a self-similar blob over the orbital sheet, area q thickened by its own radius \sqrt q.” Up (q=\tfrac23) carries more color-volume than down (q=\tfrac13); colorless objects carry none and sit on the bare pairing-two. This is a reading of the exponent, not a derivation — but the closed form itself, with its prefactor locked to the pairing-two, replaces three unexplained Koide ratios (0.667,\,0.731,\,0.845) with one formula in the charge.
What color sets, and what color frees
The \mathbb Z_3 fit \sqrt{m_k}=\bar M(1+A\cos\theta_k) is exact — three masses in, three numbers (\bar M, A, \delta) out, no residual — so reading it on each triad cleanly splits the charged-fermion mass content into two sectors. The amplitude sector (Q, how far the triad tilts from the democratic axis) is charge-set for all four by the one functional above. The phase sector (\delta, where on the circle the triad sits, which fixes the within-generation hierarchy) behaves differently:
| Triad | A/\sqrt2 | Amplitude Q | Phase 3\delta | 3\delta-Q | locked? |
|---|---|---|---|---|---|
| neutrino (colorless) | 0.62 | 0.462 | 0.809 | +0.35 | no — overshoot |
| charged lepton (colorless) | \mathbf{1.00} | 0.6667 | 0.6667 | \mathbf{0.000} | yes |
| down quark (colored) | 1.09 | 0.731 | 0.330 | -0.40 | no — undershoot |
| up quark (colored) | 1.24 | 0.845 | 0.230 | -0.61 | no — undershoot |
The naive reading — “color frees the phase” — is wrong, and the neutrino is what shows it: the neutrino is colorless yet its phase is unlocked just like the quarks’. What the lock 3\delta=Q actually tracks is sitting exactly on the pairing balance point A=\sqrt2 — and the charged lepton is the only triad there. Order the triads by amplitude and the lock is the sign change of 3\delta-Q: push the amplitude below \sqrt2 (the neutrino, dragged down by the floor) and the phase overshoots; push it above (the quarks, loaded up by color) and it undershoots; the lock is the crossing, and only the bare colorless clock at A=\sqrt2 lands on it. Color is one way off the point, the floor is the other — either departure unlocks the phase.
Two facts harden this from “an observed ordering” into structure (scripts/koide_phase_lock.py). First, an increasing curve and a decreasing curve cross exactly once: the Koide spread Q(A)=\tfrac13+A^2/6 is derivably increasing in A, while the measured phase 3\delta runs the other way (0.81\to0.67\to0.33\to0.23 down the ladder). A sign change of 3\delta-Q is therefore forced — the open question shrinks to “why does the crossing land at A=\sqrt2.” Second, at the lock the phase decomposes as \delta=Q/3=\tfrac19+A^2/18, and A=\sqrt2 is the unique amplitude where the pairing piece A^2/18 equals the democratic floor \tfrac19. So the balance point is literally where the phase’s two halves balance — turning the open target from a coordinate (“why at \sqrt2”) into an equilibrium condition (“why the lightest member equalizes its democratic and pairing halves”), the kind of statement a stress-tensor minimization could actually output.
This localizes the framework’s whole remaining debt on the charged-fermion masses to one place. The lepton triad is fully locked — amplitude charge-set (Q=2/3) and phase pinned (3\delta=Q, i.e. \delta=2/9) — which is why one anchor plus no further input lands all three lepton masses to 0.001\%. The quark triads are half-locked: their amplitudes are charge-set, but their phases carry the un-derived within-generation hierarchy. So, setting aside the overall mass scale of each triad (the one borrowed anchor apiece — the electron’s m_e=\alpha_{mf}m_\text{eff} and the analogous up/down scales of the open Yukawa program), the entire residual freedom in the ratio structure of the charged-fermion mass matrix is now just the two quark within-generation phases, and nothing else.
Mixing: why leptons mix large and quarks mix small
The same clock says something about mixing — the misalignment between the mass basis (the harmonic eigenstates) and the weak basis (which strained-boundary configuration the W couples to), read out as the CKM matrix for quarks and the PMNS matrix for neutrinos. The two matrices look nothing alike, and that contrast is one of the Standard Model’s deepest unexplained facts:
- Quarks (CKM) barely mix. The matrix is nearly the identity: the Cabibbo angle is only \theta_C\approx13^\circ, and the other two angles are tiny (V_{cb}: \sim2.3^\circ, V_{ub}: \sim0.2^\circ) — a steep hierarchy, each generation gap suppressing mixing further.
- Neutrinos (PMNS) mix hugely. \theta_{12}\approx34^\circ, \theta_{13}\approx8.5^\circ, and the atmospheric angle \theta_{23}\approx43–49^\circ, consistent with maximal 45^\circ.
In the clock picture, mixing between two triads is the relative rotation of their two \mathbb Z_3 clocks — and the lock table above already says which way each clock is turned. The quark pair is turned the same way: up and down both sit below the lock (3\delta-Q = -0.61 and -0.40), color-loaded to the same side. Two clocks rotated together stay nearly aligned, so their bases nearly cancel — small CKM. The lepton pair straddles the lock: the charged lepton sits on it (3\delta-Q=0) while the neutrino is dragged the other way by the visibility floor (+0.35). Two clocks turned opposite ways are maximally misaligned — large PMNS. This is the concrete form of the framework’s standing guess (WIP-29): color co-loads both quark sectors identically and so aligns their two bases, while the leptons carry no common loading. The mechanism fixes the sign — quarks co-loaded and near-diagonal, leptons straddling and large — cleanly; it does not by itself set the angles (the raw 3\delta-Q gaps, 0.21 for quarks versus 0.35 for leptons, run the right direction but understate how dramatically CKM collapses). The obvious quantitative target here — whether the clock reproduces the Gatto–Sartori–Tonin relation \theta_C\approx\sqrt{m_d/m_s} — was checked and comes back honest but partial (scripts/wip29_gst_cabibbo_check.py): the square-root form is native, since the clock’s own variable is \sqrt m, so the ratio of two adjacent clock amplitudes is \sqrt{m_\text{light}/m_\text{heavy}} identically, no square root applied by hand. But that fixes only the shape; the angle’s magnitude, the up/down pairing, and the relative sector phase remain empirical inputs, not clock outputs — so it does not lift the claim above “sign only,” and we do not present it as a derivation of the Cabibbo angle.
A suggestive anchor: maximal atmospheric mixing sits at the pairing-\sqrt2 value. The most robust number in the lepton sector is \theta_{23}\approx45^\circ, and the framework already owns a 45^\circ: the democratic-to-hierarchy tilt of the Koide vector, \cos45^\circ=1/\sqrt2, the pairing-two written as an angle. A bare neutrino clock — junction-free, twist-free, carrying neither color nor charge to push it off balance — has nothing to move it away from that pairing point, so the two lepton clocks plausibly meet at the 45^\circ that is the maximal-mixing angle, while dressed quarks, loaded off the point by color, cannot reach it (the same statement as “small CKM,” once more). Honesty check on how much this buys: the two \sqrt2’s being identified live in different spaces — A=\sqrt2 is an amplitude in \sqrt{\text{mass}} space (how far a triad leans off the democratic axis), whereas \cos45^\circ=1/\sqrt2 is a rotation angle between two flavor bases. That both come out \sqrt2 is the same re-encoding move as the trimaximal reading below, not a tighter or independent result — a numerical coincidence the framework can absorb naturally, not a mechanism that forces 45^\circ from a boundary calculation.
A structural reading (Koide-grade, not a forward number). The canonical unitary of \mathbb Z_3 is the 3\times3 discrete-Fourier (“magic”) matrix, and its \mathbb Z_3-singlet combination is the democratic column (1,1,1)/\sqrt3 — every flavor mixed equally, |U_{\alpha i}|^2=\tfrac13. A democratic PMNS column is exactly trimaximal mixing, a measured, data-consistent feature of the lepton matrix (it predicts \sin^2\theta_{12}=1/(3\cos^2\theta_{13})\approx0.34 against the measured 0.31). So the same \mathbb Z_3 that forces “three generations” and Koide’s 2/3 carries, as its natural mixing structure, the democratic lepton mixing we see — while CKM is that same magic matrix with the democracy broken by the common color-loading above, collapsing it toward the identity. This is a re-encoding of a known structure through the \mathbb Z_3 junction, at the same epistemic level as the Koide result itself — a unifying picture (one object → democratic leptons and near-diagonal quarks), not a prediction of a previously-unknown angle. Notably, the discrete flavor symmetries mainstream model-builders reach for to get tri-bimaximal mixing (A_4, S_4) contain this same \mathbb Z_3 as their generating subgroup: the substrate’s three-fold junction is already the standard toolkit’s core.
CP violation is the imaginary part the clock already carries. The one place the two sectors agree — where PMNS and CKM might both be large — is the CP-violating phase, and the clock has a clean structural reason for it. A \mathbb Z_3 clock is not a real angle on a line; it is built from the cube roots of unity \omega=e^{2\pi i/3}, and its canonical mixing unitary — the discrete-Fourier “magic” matrix invoked above — has complex entries whose intrinsic phase is 2\pi/3, the maximum a three-fold object can carry. So in this framework a CP-violating phase is not a separate ingredient anyone has to reach for: it is the imaginary part of the same \mathbb Z_3 object that already fixes the generation count and the mixing structure. A vanishing CP phase would be the surprise, requiring the imaginary part to cancel; a large one is the default. This reframes the CKM/PMNS contrast one level deeper. What differs between the two sectors is only the real mixing angles — collapsed toward the identity for the color-co-loaded quarks, left large for the straddling leptons — not the phase, which both sectors inherit near-maximal from the shared complex unitary. The data is consistent with exactly this split: the Dirac phase is O(1) in both sectors (\delta_\text{CKM}\approx65^\circ; the current PMNS hint sits near maximal, \delta_\text{PMNS}\sim-90^\circ), and the CKM effect looks tiny only because the Jarlskog invariant multiplies that large phase by the small quark angles — J_\text{CKM}\sim3\times10^{-5} is small because \sin\theta_C is, not because \sin\delta is. The unified statement: one complex \mathbb Z_3 clock supplies a large CP phase to both sectors; the sectors differ only in how far color collapses their real angles. Like the trimaximal reading, this is a re-encoding at Koide-grade — it explains why a large CP phase is generic rather than fine-tuned, and predicts the phase is near-maximal in the lepton sector, but it does not forward-compute the phase’s numerical value from a boundary calculation.
Predictions and falsification
- Koide Q=2/3 is structural, not accidental. Forced by the three-fold junction plus the pairing-\sqrt2. As the lepton masses sharpen (chiefly m_\tau), Q must stay pinned at 2/3; a robust drift away at the few-ppm level would falsify the \mathbb Z_3+\sqrt2 reading. Current value 0.666660 is consistent.
- The phase is \delta\simeq2/9 rad. A sharpened m_\tau that pulls the best-fit phase decisively away from 2/9 would break the bet while leaving the Q=2/3 core intact — the two claims are separable by design.
- Exactly three. A fourth sequential charged-lepton generation sharing this structure is forbidden on the same ground as a fourth stable junction valence — consistent with LEP’s N_\nu=2.984\pm0.008 and with direct fourth-generation searches.
- The quark Q’s are the charge-set values Q=(2+q^{3/2})/3. The color-loading functional predicts Q_\text{down}=0.731 and Q_\text{up}=0.848 from the electric charge alone (measured 0.731,\,0.845). Because Q is invariant under flavor-universal rescaling, no choice of common scale brings the quarks back to 2/3; they must stay on the charge-set values, not drift to generic numbers.
- The neutrino offset is the floor’s. Its colorless base is 2/3, so Q_\nu should sit below 2/3, dragged toward 1/3 by the floor — prediction Q_\nu\approx0.46, capped below the 0.585 a massless lightest neutrino would give. A measured Q_\nu above \sim0.585 would require the lightest mass below the framework’s floor, which the framework forbids.
- Mixing follows the lock, not the mass. The clock reading of mixing predicts the sign: quark (CKM) mixing stays small and hierarchical because up and down are color-loaded to the same side of the lock, while lepton (PMNS) mixing stays large because the charged lepton and the floor-dragged neutrino straddle it — with the atmospheric angle pinned near the pairing-\sqrt2 value \theta_{23}\approx45^\circ. A future determination driving \theta_{23} decisively away from maximal, or a fourth generation, would strain the reading; the CKM-small / PMNS-large contrast is the firm qualitative claim.
- CP violation is generic and near-maximal in the lepton sector. Because the \mathbb Z_3 mixing unitary is built from the complex cube roots of unity, a large Dirac CP phase is the framework’s default, not a tuned add-on — the same phase in both sectors, with only the real angles differing. The falsifiable edge is the leptons: the PMNS Dirac phase should land near maximal (|\delta_\text{PMNS}|\sim90^\circ). A future determination pinning \delta_\text{PMNS} decisively near 0^\circ or 180^\circ (CP-conserving) would contradict the “imaginary part is always present” reading, while leaving the real-angle mechanism (bullets 1–6) intact.
Honest assessment
Genuinely predicted, zero parameters: Koide’s Q=2/3 and “why three.” Both follow from two structures already in the paper — the \mathbb Z_3 three-fold of the junction and the pairing-\sqrt2 — with no freedom left, matching at 9 ppm. The framework did not reach for Koide; the 2/3 fell out of the same three-fold that already gave color and the quark charges. This is the solid core, and it does not depend on the bet.
The bet: that the residual phase is exactly \delta=2/9 rad — most economically the cross-sector identity 3\delta=Q, framed by the massless edge \delta=\pi/12 that is derived from A=\sqrt2. This is the high-risk gamble: motivated (2/9=2/3^2 is built from the same two-and-three, decomposes as democratic 1/9 plus pairing A^2/18, and is the value Koide’s own analysis flags), but 3\delta=Q is not yet derived from a boundary calculation. What makes it more than a one-knob fit is the over-determination — a single angle lands two independent ratios to 0.001\% and 0.007\% at once, and the data hands back a simple rational rather than a generic number.
Borrowed: the overall scale \bar M — the electron mass itself, set by the existing identity m_e=\alpha_{mf}m_\text{eff}. This chapter fixes the ratios; the scale is the anchor the paper already carries.
The strongest thing here is convergence. The paper did not invent a mechanism for the generations. It noticed that the three-fold junction it had already built — on the strong-force and charge side, for entirely separate reasons — is the same \mathbb Z_3 whose cube-roots-of-unity structure makes Koide’s 2/3 unavoidable, and that the pairing-\sqrt2 it had already built — for the lattice and the superconductors — is the one amplitude that turns “2/N” into “2/3.” The deepest unexplained numbers in the fermion sector, why three and m_\mu/m_e, turn out to be two faces of one geometric object the framework keeps returning to: a knot with three branches, turning.
Shared status with the 6\pi^5 proton-ratio result. Worth recording plainly, because the two are built from the same three-fold junction and it fixes what “prediction” means here: this chapter and the 6\pi^5 breadcrumb sit at the same epistemic level. In both, a known empirical relation — Koide’s Q=2/3 (1981) here, Lenz’s m_p/m_e\approx6\pi^5 (1951) there — is given a structural account via the \mathbb Z_3 junction; in both, the “exact” value the geometry forces (exactly 2/3; exactly 6\pi^5) misses the measured value by a small unexplained residual of the same shape (the {\sim}6–9 ppm by which the real leptons fall below 2/3; the 19 ppm by which m_p/m_e sits above 6\pi^5); and in both, the remaining precision is carried by an admittedly-underived number (the identity 3\delta=Q here, the finite level k there). So the honest claim in each is a re-encoding of a pre-existing coincidence into the framework’s geometry — a genuine unifying picture, and in the lepton case an over-determined one (one angle, two ratios), but not a forward prediction of a previously-unknown number. A focused 2026-07 search for a genuinely new number the same machinery forecasts rather than reproduces turned up none; that, not the ppm-level hits, is the honest boundary of the program.
Putting the section in context
The Standard Model chapter and the generation puzzle built the structure of the fermion families — radial harmonics, chirality sheets, a three-generation ceiling — and left the numbers as the framework’s largest open debt on the Standard Model side. This chapter pays the structural part of that debt: the count (three) and the cleanest ratio relation (Koide’s 2/3) come out with no free parameter, from the three-fold junction the paper already uses everywhere, and the charged-lepton masses follow from one residual phase the data hands back as the rational 2/9. With the neutrino floor pinning the bottom of the ladder and the \mathbb Z_3+\sqrt2 structure setting the rungs, the fermion mass spectrum — once thirteen free Yukawa couplings — reads as the electron anchor, one three-fold, one pairing-two, and one phase. The remaining open piece is to derive that phase, the same boundary computation the Yukawa program has been pointing at all along.