REVIEW 5 minor 6 references
Unit-conjugacy invariants of a Hurwitz prime cannot determine the labelled metacommutation permutation, or even the destination of one class; only its cycle shape is fixed.
Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →
T0 review · deepseek-v4-flash
2026-08-05 00:11 UTC pith:JNNRHCCQ
load-bearing objection A clean, explicit negative result that closes the invariant-theoretic boundary for metacommutation; deserves a serious referee.
Conjugation invariants determine the metacommutation permutation only up to relabelling
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
The paper's central claim is that metacommutation permutations are determined by conjugation-invariant data of the prime Q only up to relabelling: no function I(Q) invariant under uQu^{-1} can recover the labelled permutation π_Q of the p+1 classes of norm-p primes, nor even the image of one specified class. The proof exhibits the minimal case (p,q)=(3,5): the four Hurwitz primes 2+i, 2+j, 2+k, and 2-i lie on one unit-conjugacy orbit, so every invariant I agrees on them, yet their metacommutation actions are four pairwise distinct 4-cycles on the same four classes, already differing at the first class. Thus the invariant formulas for sign, fixed points, and cycle length are complete: they co
What carries the argument
The central object is the metacommutation permutation π_Q, defined by uniquely refactoring PQ=Q'P' for primes P,Q of distinct norms. The load-bearing identity is the equivariance relation π_{uQu^{-1}} = ρ_u π_Q ρ_u^{-1}, where ρ_u is conjugation of classes by the unit u; this converts an active unit conjugation into conjugation inside the symmetric group, so whenever the conjugates of Q give different permutations, invariant data cannot determine the labelled map. The explicit witness at (3,5) supplies four primes in one unit-conjugacy orbit whose induced permutations are pairwise distinct, and the frame torsor plus orbit-stabilizer coset pricing complete the boundary.
Load-bearing premise
The witness rests on the sixteen explicit refactorizations in Appendix A being correct with the stated right factors; one sign error or swapped factor would collapse the pairwise distinction of the four cycles.
What would settle it
Expand every identity in Appendix A exactly, by hand or with exact integer arithmetic, and verify the left side equals the stated product and the right factor is the listed norm-3 representative; then check that the four permutations on (P0,P1,P2,P3) are pairwise distinct and already differ at P0.
If this is right
- The invariant layer is complete: the known sign, fixed-point, and cycle-length formulas are all that any conjugation-invariant function of Q can determine.
- At (p,q)=(3,5), four primes that every conjugation invariant sees as identical must be treated as different by any procedure that outputs a destination; evaluating the action of Q is unavoidable.
- No canonical projective labelling of the p+1 classes exists: frames form a torsor under PGL2(Fp), so the algebra prefers no labelling, not even of three classes.
- The cost accounting is exact: one named destination equals one coset g_Q G_C in PGL2(Fp)/G_C, and the full labelled permutation equals the acting element g_Q itself.
Where Pith is reading between the lines
- The same equivariance-plus-witness recipe should be testable at other prime pairs: find a unit-conjugacy orbit of norm-q primes whose induced permutations on norm-p classes are pairwise distinct, and the invariant-blindness proof repeats verbatim; the paper only demonstrates the minimal case (3,5).
- A natural reading is that any deterministic procedure that only queries conjugation-invariant data of Q cannot compute destinations, even though the paper explicitly disclaims computational lower bounds; with a fixed frame, every destination is cheaply computable by the Mobius rule.
- The paper's shape/step/casting separation is a concrete instance of a general distinction between unlabelled and labelled structures: invariant statistics can fix an isomorphism class while the named object needs extra choices, here a frame.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies the metacommutation permutation π_Q induced on the p+1 classes of Hurwitz primes of norm p by a prime Q of norm q. It proves that conjugation-invariant data of Q, however complete as a class function, cannot determine π_Q as a labelled permutation of the intrinsic class set, and cannot even determine the image of one fixed class. The proof uses an equivariance identity (Lemma 2) and a fully explicit witness at (p,q)=(3,5): four norm-5 primes that lie in one unit-conjugacy orbit but induce four pairwise distinct 4-cycles on the same four classes. It also establishes that isomorphisms from H/pH to M_2(F_p) form a torsor under PGL_2(F_p), so no projective labelling of the classes is canonical, and that the datum of one destination π_Q(C) is exactly a coset of the stabilizer of C.
Significance. If correct, the result cleanly delineates the boundary of the earlier formulas of Cohn--Kumar and Leite--Machiavelo: cycle type is the full information that conjugation invariants can carry, while the labelled permutation and even single-step destinations require strictly more. The paper's main strength is its verifiability: the sixteen refactorizations in Appendix A are explicit and each can be checked by hand, and Theorem 1 does not depend on the motivational Section 6 or the unpublished monograph [5]. The paper is also honest about what it does not prove: it gives no computational lower bound and notes that the equivariance identity is elementary. This is a focused, sound contribution.
minor comments (5)
- [Abstract and Section 3] The phrase 'the four primes ... form a single unit-conjugacy orbit' is slightly stronger than what is proved and used. Those four primes do lie in a single orbit, but the full orbit also contains other norm-5 primes (e.g. 2−j and 2−k). Rephrase as 'lie in a single unit-conjugacy orbit' to avoid the implication of exhaustion.
- [Section 2] The notation H^\times is introduced as 'the 24 Hurwitz units'. In quaternion contexts H^\times usually denotes the nonzero elements. Consider using H^1 or U(H) for the unit group, or add an explicit sentence to avoid ambiguity.
- [Section 6] The drawn/geometric motivation is explicitly not used in the proof and depends on an in-preparation monograph [5]. It is fine as motivation, but it could be moved to a clearly marked remark or shortened, and the figures should be checked in the final version. This is not a correctness issue.
- [Appendix A] The sentence 'Each line ... verifiable by expanding the single quaternion product on the right' is slightly imprecise: to verify an identity one expands the product on the right and compares with the displayed middle term, and also checks the left equality. It would help to state the quaternion multiplication convention (ij=k, ji=-k) at the start of the appendix.
- [Section 5] When defining G_p := (H/pH)^\times / F_p^\times, it would be clearer to state that F_p^\times is embedded as central scalars, and that the quotient is taken in the usual way. This is a minor readability point.
Circularity Check
No significant circularity; only a non-load-bearing self-citation to the author's in-preparation monograph.
full rationale
Theorem 1's proof is self-contained: Lemma 2 derives equivariance directly from the defining metacommutation factorization, and the counterexample rests on explicit quaternion identities in Appendix A, each checkable by expanding one product. The four norm-5 primes are shown to lie in one unit-conjugacy orbit by explicit conjugations by rho and j, so any conjugation-invariant I is constant on them; the Appendix then exhibits four pairwise-distinct 4-cycles. None of these steps reuses the theorem's conclusion or fits a parameter and calls it a prediction. The only self-citation, [5], appears in Section 6, which is explicitly labeled 'motivation rather than proof' and is not used to establish Theorem 1, Proposition 6, or Proposition 7. Prior results [2,4,6] are cited as context and consistency checks, not as load-bearing inputs. Thus no circular step is present; the minor self-reference warrants at most a score of 2.
Axiom & Free-Parameter Ledger
axioms (5)
- domain assumption H/pH is a central simple F_p-algebra isomorphic to M_2(F_p) for odd prime p
- standard math Skolem-Noether theorem: every automorphism of M_2(F_p) is inner
- domain assumption Metacommutation is the Mobius action of [\bar Q] in PGL_2(F_p), from Forsyth-Gurev-Shrima [4]
- domain assumption Cohn-Kumar formulas for sign and fixed-point count, and Leite-Machiavelo cycle structure
- standard math Orbit-stabilizer theorem: the orbit map G_p -> C_p induces a bijection G_p/G_C -> C_p
Cite this review
Pith. "Pith review of Conjugation invariants determine the metacommutation permutation only up to relabelling." pith.science (2026). https://pith.science/paper/JNNRHCCQ
@misc{pith2026260801610,
author = {Pith},
title = {Pith review of: Conjugation invariants determine the metacommutation permutation only up to relabelling},
year = {2026},
howpublished = {\url{https://pith.science/paper/JNNRHCCQ}},
note = {Machine review of arXiv:2608.01610}
}
read the original abstract
Let $\mathcal{H}$ be the Hurwitz quaternions, $p$ an odd prime, and $Q \in \mathcal{H}$ a prime of norm $q \neq p$. Metacommutation $PQ = Q'P'$ induces a permutation $\pi_Q$ of the $p+1$ left-associate classes of primes of norm $p$. Cohn and Kumar compute its sign and fixed-point count, and Leite and Machiavelo its full cycle structure, by formulas depending only on conjugation-invariant data of $Q$ (namely $q$ and $\mathrm{tr}\,Q$). We prove this is exactly the boundary of what such invariants can carry: no quantity $I(Q)$ invariant under unit conjugation determines $\pi_Q$ as a labelled permutation of the intrinsic class set, or even the image of a single specified class. The proof combines an equivariance identity $\pi_{uQu^{-1}} = \rho_u \pi_Q \rho_u^{-1}$ with a minimal, fully explicit witness at $(p,q) = (3,5)$: the four primes $2+i$, $2+j$, $2+k$, $2-i$ form a single unit-conjugacy orbit, hence agree under every conjugation-invariant function, yet induce four pairwise distinct $4$-cycles of the same four classes. We further observe that isomorphisms $\mathcal{H}/p\mathcal{H} \to M_2(\mathbb{F}_p)$ form a torsor under $\mathrm{PGL}_2(\mathbb{F}_p)$, so no projective labelling of the classes is canonical, and that by orbit-stabilizer the datum of one destination $\pi_Q(C)$ is exactly a coset $g_Q G_C$ in $\mathrm{PGL}_2(\mathbb{F}_p)/G_C$. All sixteen refactorizations in the witness are listed in the appendix and have been verified by machine along two independent routes.
Figures
Reference graph
Works this paper leans on
-
[1]
Metacommutation of primes in Eichler orders
A. Babei and S. Chari,Metacommutation of primes in Eichler orders, arXiv:1909.12915
work page internal anchor Pith review Pith/arXiv arXiv 1909
-
[2]
H. Cohn and A. Kumar,Metacommutation of Hurwitz primes, Proc. Amer. Math. Soc.143(2015), 1459– 1469
work page 2015
-
[3]
J. H. Conway and D. A. Smith,On Quaternions and Octonions: Their Geometry, Arithmetic, and Symmetry, A K Peters, 2003
2003
-
[4]
Metacommutation as a Group Action on the Projective Line over $\mathbf{F}_\mathbf{p}$
A. Forsyth, J. Gurev, and S. Shrima,Metacommutation as a group action on the projective line overF p, arXiv:1503.06259
work page internal anchor Pith review Pith/arXiv arXiv
-
[5]
Fried,Five-Eighths: A proof-by-picture monograph on the extremal groups, in preparation
M. Fried,Five-Eighths: A proof-by-picture monograph on the extremal groups, in preparation
-
[6]
On the Cycle Structure of the Metacommutation Map
A. Leite and A. Machiavelo,On the cycle structure of the metacommutation map, arXiv:2504.08709
work page internal anchor Pith review Pith/arXiv arXiv
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.