Submitted:
15 September 2026
Posted:
17 September 2026
You are already at the latest version
Abstract
We separate two questions. (Q1, empirical) Do the pole masses satisfy the Koide relation in the polar form \(\sqrt{m_i}=\mu(1+\sqrt2\cos(\varphi_0+2\pi i/3))\) with \(\varphi_0=2\pi/3+2/9\)? No: the same-scheme comparison of \(m_\mu/m_e\) gives 206.77032 vs. 206.7682830(44), i.e. \(\approx460\sigma\), with no theoretical error. (Q2, fundamental) Can a UV theory derive it? The obstruction is scheme dependence and it is large: the common-scale MS ratios differ from the pole ratios by 2.1% (\(m_\mu/m_e\)) and 3.2% (\(m_\tau/m_e\)), moving Q by \( +1.26\times10^{-3}=185\sigma \) and \( \varphi_0 \) by \( -1.19\times10^{-3} \) rad. As the charged-lepton mass anomalous dimension is flavour-universal, there is no continuous scheme dial: only the pole point and the (common-scale) MS point. Hence any UV derivation must explain why the relation holds for pole masses, and no exact \(\varphi_0\) claim is scheme-invariant (\(\Delta\varphi_0=6800\times|\varphi^*-2/9|\)). We also prove an unconditional theorem: for any rational mass triplet, the fitted phase \(\varphi^*\) is a transcendental number of radians — so \(\varphi^*=2/9\), \(1/3\), \(\sqrt2/9\), or any nonzero algebraic radian, is exactly excluded, independently of errors and schemes. Finally the conditional pole-scheme prediction \(m_\tau=1776.969\) MeV (\( 0.85\sigma \)).

Keywords:
Koide relation
; charged-lepton masses
; pole masses
; renormalization scheme
; transcendental number theory
1. Introduction
The charged-lepton spectrum is the most precisely measured part of the fermion sector, and it is also the least structured: the three pole masses carry no continuous quantum numbers, so any exact relation among them is either an accident of the flavour sector or a clue to its dynamics. In 1982, within a composite fermion–boson model of quarks and leptons, Koide observed [1,2,13] that the charged-lepton masses satisfy
a relation carrying no free parameters, satisfied by the measured pole masses at the level of (; §Section 3). It is not obeyed by the quark or neutrino spectra [7,20], although Koide-type relations have been constructed for them [21]; why the relation singles out the charged leptons remains part of the puzzle.
The relation has a transparent geometric content. As first noted by Foot [3] and emphasised subsequently by Brannen [4], the three can be written as the projections of a single vector onto a cone, ; then is exactly , and the spectrum is described by one shape angle . Empirically ( rad ), and both the geometric reading and the interpretation of this phase have attracted a sizeable literature [19,22]. Because the polar form is a discrete Fourier transform of the , the relation appears natural in models with a discrete flavour symmetry, where and are fixed by vacuum alignment.
Attempts to derive or explain the relation form a substantial literature and, judged by outcome, remain inconclusive. Following Koide’s own family-symmetry constructions [14], concrete models have been based on lepton family symmetries [18], on a family gauge symmetry in which holds already at tree level and survives radiative corrections [17], on Yukawaon models [6], and on supersymmetric variants that stabilise the relation against renormalisation-group running [8,16]. Koide himself has described the origin of the formula as an open “mystery” [15], and the relation has been used as a quantitative constraint on flavour and renormalisation-group scenarios well beyond its original setting [9,23].
A feature common to essentially all of these proposals is that they predict masses in a scheme: a UV model naturally delivers running masses at some renormalisation scale, whereas the empirical relation is stated for, and satisfied by, pole masses. The scale dependence of Koide-like relations was quantified by Xing and Zhang [7], but its consequences for model building — and for the status of the phase — have not been drawn out.
This note separates the empirical question from the fundamental one and states three results, each with its precise status. (i) An unconditional theorem (§4): for any rational mass triplet the fitted phase is a transcendental number of radians, so — or any nonzero algebraic number of radians — is excluded exactly, independently of errors and of any scheme. (ii) A quantitative scheme analysis (§5) showing that the relation is a pole-mass statement: the common-scale ratios differ from the pole ratios by 2–, moving Q by . (iii) A conditional pole-scheme prediction (§7): MeV . The rest of the note is organised accordingly.
2. The Parametrization
3. The Empirical Question (Q1)
Table 1 lists the inputs (pole values reproducing those of the original Koide literature [1,2]; see caption for provenance). M0 predicts (from and alone, no fitting) , to be compared with the pole measurement [11]: a deviation of , i.e. , with no theoretical error (both are pole-mass quantities). Within the pole scheme the phase is likewise pinned, , so rad and M0 is excluded at the same significance.
4. An Unconditional Theorem: The Fitted Phase Is Transcendental
Theorem.Let (every measured value is a terminating decimal) and let be the phase of the polar fit. Then
lies in , hence is algebraic. Since , is transcendental (in radians). Proof. The are algebraic, so is algebraic; if were a nonzero algebraic number of radians, Lindemann–Weierstrass [12] would make transcendental and hence transcendental — contradiction. ▪ Numerically (50-digit mpmath; integer minimal polynomial via sympy), satisfies an eighth-degree integer polynomial. Corollary. , and in fact is not any nonzero algebraic number of radians — exactly, unconditionally, independently of the error bars or the scheme. Equivalently (contrapositive): a geometric model can predict only or a rational multiple of ; never a rational number of radians. (The analogous statement holds for any finite algebraic-coordinate point set: any nonzero inter-vector angle in radians is transcendental.)
5. The Fundamental Question (Q2): Koide Is a Pole-Mass Statement
A UV theory predicts masses at some scale; the pole masses follow by running and matching. Using the running masses of Xing and Zhang [7] (Table 1),
| pole | () | change | |
| Q | () | ||
| r | |||
| rad |
The experimental standard deviation of Q is ; the shift is (and the empirical deviation ).
No continuous dial. In , with flavour-universal ( identical; broken only by ), so the ratios are scale-independent. There are thus three scheme points, not a continuum: pole, at a common scale, and at each lepton’s own scale — the last differing from pole by only , so holds there too. The Koide relation therefore holds for the pole masses and for each-at-own-scale (differing by ), but fails by at a common scale.
Corollary 1. is a pure observable (no theoretical error). If a UV theory predicts in , matching to pole masses gives , excluded at . Any UV derivation of the Koide relation must therefore explain why it holds for pole masses [7,8,9]. This is a real constraint on model building.
- Corollary 2 (testability no-go). moves across schemes by rad, . No bare numerical claim about — including — is scheme-invariant.
- Corollary 3. Once the pole scheme is fixed (in which the relation is meaningful), the data give , so is excluded at ; independently, the theorem of §Section 4 excludes it exactly. “Scheme-contingent” (Corollary 2) and “excluded” (here) apply to different statements.
6. Not a Discrete-Symmetry Phase (Pole Scheme)
For residual flavour symmetries one expects with small q. In the pole scheme the window is rad, giving (with ): , , , , . So is not a small-denominator rational multiple of in the pole scheme. (Across schemes the window is in and even falls to ; since UV models predict quantities, this is an observation about the pole-scheme value, not a constraint on model building.)
7. The Conditional Prediction
Fixing , has the closed form ()
With pole inputs this is from MeV (combined ; the input error from is MeV and negligible). The prediction is conditional on the relation holding for pole masses: applying to the masses instead gives MeV vs. the actual MeV — off by 22 MeV. FCC-ee / CEPC threshold scans ( MeV) would test the pole-scheme version at . Present-day single measurements (e.g. BaBar threshold scans vs. Belle) spread over MeV [10], comparable to the offset, so no present anomaly is claimed.
8. Outlook
Any UV explanation must target Q in the pole scheme and explain the pole-mass selection; the phase direction is closed in three independent ways — exact transcendence (§4), scheme non-invariance (Corollary 2), and the pole-scheme exclusion at (Corollary 3). We note the scale/scheme dependence of the Koide relation is a known issue [7,8,9].
Priority & method. All numbers in Tables and §4 are reproduced by a single 50-digit mpmath script (with sympy for the minimal polynomial); input provenance is stated in Table 1. The scale-dependence observation is due to Xing and Zhang [7]; we thank an anonymous referee. Generative-AI assistance was used for analysis and drafting.
Supplementary Materials
The following supporting information can be downloaded at the website of this paper posted on Preprints.org.
Data Availability Statement
No data were used for the research described in the article. All numerical results are analytic and are reproduced by the script described in the text.
Conflicts of Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References
- Koide,, Y. Lett. Nuovo Cim. 1982, 34, 201.
- Koide, Y. Phys. Rev. D. 1983, 28, 252.
- Foot, R. A note on Koide’s lepton mass relation. arXiv arXiv:hep.
- Brannen, C. A. The lepton masses (2006), unpublished note, http://brannenworks.com/MASSES2.pdf. In Ann. Rev. Nucl. Part. Sci.; Mohapatra, R. N., Smirnov, A. Y., Eds.; 2006; Volume 56, p. 569. [Google Scholar]
- Koide, Y. Tribimaximal neutrino mixing and a relation between neutrino- and charged lepton-mass spectra. J. Phys. G. 2007, 34, 1653. [Google Scholar] [CrossRef]
- Koide, Y.; Nishiura, H. Universal bilinear form of quark and lepton mass matrices. Phys. Rev. D. 2014, 90, 016009. [Google Scholar] [CrossRef]
- Xing, Z.-z.; Zhang, H. On the Koide-like relations for the running masses of charged leptons, neutrinos and quarks. Phys. Lett. B 2006, 635, 107. [Google Scholar] [CrossRef]
- Koide, Y.; Yamashita, T. Charged lepton mass relations in a SUSY scenario. Phys. Lett. B 2018, 787, 171. [Google Scholar] [CrossRef]
- LeClair, A. A rich structure of renormalization group flows for Higgs-like models in 4 dimensions. arXiv arXiv:2411.07476.
- Workman, R. L.; et al. (Particle Data Group). Prog. Theor. Exp. Phys. 2022, 2022, 083C01. [Google Scholar]
- Tiesinga, E. Rev. Mod. Phys. 2021, 93, 025010. [PubMed]
- Baker,, A. Baker, Transcendental Number Theory; Cambridge Univ. Press, 1975. [Google Scholar]
- Koide, Y. A fermion–boson composite model of quarks and leptons. Phys. Lett. B 1983, 120, 161. [Google Scholar] [CrossRef]
- Koide, Y. Universal texture of quark and lepton mass matrices with an extended flavor 2↔3 symmetry. Phys. Rev. D. 2004, 69, 093001. [Google Scholar] [CrossRef]
- Koide, Y. Challenge to the mystery of the charged lepton mass formula. Proc. Lepton-Photon 2005, arXiv:hep. [Google Scholar] [CrossRef]
- Koide, Y. Charged lepton mass relations in a supersymmetric Yukawaon model. Phys. Rev. D. 2009, 79, 033009. [Google Scholar] [CrossRef]
- Sumino, Y. Family gauge symmetry as an origin of Koide’s mass formula and charged lepton spectrum. JHEP 2009, 05, 075. [Google Scholar] [CrossRef]
- Ma, E. Lepton family symmetry and possible application to the Koide mass formula. Phys. Lett. B 2007, 649, 287. [Google Scholar] [CrossRef]
- Rivero, A.; Gsponer, A. The strange formula of Dr. Koide. arXiv arXiv:hep.
- Kartavtsev, A. A remark on the Koide relation for quarks. arXiv arXiv:1111.0480.
- Rivero, A. A new Koide tuple: strange-charm-bottom. arXiv arXiv:1111.7232.
- Kocik, J. The Koide lepton mass formula and geometry of circles. arXiv arXiv:1201.2067.
- Rivero, A. New sum rules of the Koide type. Phys. Lett. B 2026, 877, 140510. [Google Scholar] [CrossRef]
Table 1.
Mass inputs. Pole values from Xing–Zhang [7], Eq. (1) (PDG 2004 era); the current PDG/CODATA values [10,11] differ in the 8th digit (, immaterial to the conclusions). values at from [7], Eq. (8). The last column is an independent one-loop QED running estimate, agreeing to .
| pole (MeV) | at (MeV) | cross-check | |
|---|---|---|---|
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2026 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Copyright: This open access article is published under a Creative Commons CC BY 4.0 license, which permit the free download, distribution, and reuse, provided that the author and preprint are cited in any reuse.