REVIEW 3 major objections 5 minor 11 references
The range of the des statistic for conjugacy classes in $S_n$
T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash
Pith's one-line read In every non-identity conjugacy class of $S_n$, descent counts take all values from 1 to the class maximum.
desk verdict The interval theorem is plausible and new, but a false left/right multiplication lemma in Proposition 3.5 breaks the submitted proof; the paper is worth a revision, not as-is. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The proof is carried by two mechanisms. First is the local classification of Proposition 3.2: conjugating a permutation by the adjacent transposition $s_i = (i, i+1)$ swaps the values $i$ and $i+1$ and the entries in positions $i$ and $i+1$, and the way those two swaps overlap falls into three cases; from this, Corollary 3.8 bounds the change in descent count by 2 in any single conjugation step. Second is the long cycle $w = (1,\dots,n)$: conjugation by powers of $w$ cyclically rotates the permutation, and Proposition 3.14 determines when the cyclic descent set coincides with the ordinary descent set. That coincidence is exactly what lets the proof repair a jump of size 2 by rotating the lower or upper endpoint of the jump, producing the missing descent count.
What would settle it
Run a direct enumeration for all permutations in $S_n$ for $n \le 8$: for every conjugacy class, list the set of attained descent counts and check that it equals $\{1,\dots,M\}$; any gap would refute the theorem. Separately, enumerate all pairs $(\pi, i)$ and check whether $|\mathrm{des}(\pi) - \mathrm{des}(s_i^{-1}\pi s_i)|$ ever exceeds 2; a single such pair would refute the local bound that the gap-closing argument needs.
Extended reading notes
Core claim
On the paper's own terms, the central discovery is Theorem 3.4: for any conjugacy class $C_n$ of $S_n$, writing $M = \max_{\pi \in C_n} \mathrm{des}(\pi)$, the set of attained descent counts is exactly $\{1,2,\dots,M\}$ (and $\{0\}$ for the identity class). The proof starts with a Young-diagram filling that places one descent in every non-identity class, then moves to a permutation of maximal descent count along a chain of conjugations by adjacent transpositions. Along the chain the descent count can change by at most 2, and whenever a step jumps upward by 2, one of two cyclic rotations of the endpoints yields a permutation in the same class whose descent count is exactly the missing intermediate value.
Load-bearing premise
The load-bearing premise is that conjugating by a single adjacent transposition changes the descent count by at most 2; the proof of that local bound contains a left-versus-right multiplication error, and the cyclic-descent lemmas used to close the size-2 jumps also contain indexing and notation errors, so the bound and the gap-closing step are not fully verified as printed.
Editorial extensions
If this is right
- Every non-identity conjugacy class contains at least one permutation with exactly one descent.
- The descent counts in a class form the contiguous interval $\{1,\dots,M\}$, so no value between the minimum and maximum is missed.
- The proof is constructive: for any attainable descent value in a class, a permutation with that value and the right cycle type can be produced explicitly.
- For the class of $n$-cycles, the number of minimal-descent permutations is the number of Lyndon words of length $n$, and the interval theorem implies this number is positive for every $n$.
- The range of the descent statistic on a class is determined solely by its maximum, and each class's descent distribution has support equal to that interval.
Reading between the lines
- A natural next test is whether the same interval property holds for conjugacy classes in other Coxeter groups, using simple-reflection conjugation and an analogue of the cyclic rotation; the local bound and gap-closing strategy might transfer.
- The constructive path could be turned into an algorithm that, given a cycle type and a target $k$, outputs a permutation of that cycle type with exactly $k$ descents; checking it on small $n$ would be a quick computational verification of the theorem.
- If the local bound of 2 is sharp as the paper's examples suggest, then the descent graph on each class has diameter governed by these 2-steps, and the interval property is the optimal continuity statement one could hope for.
- The contiguous-support result removes one possible source of gaps in asymptotic studies of descents over random permutations with a fixed cycle type, which may simplify normal-approximation arguments.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies the descent statistic des on conjugacy classes of the symmetric group S_n. Its main result (Theorem 3.4) states that for every non-identity conjugacy class C, the set of descent numbers realized in C is exactly the interval {1,...,M}, where M is the maximum of des on C. Theorem 2.1 proves that the minimum value is 1, via an explicit Young-diagram construction. The proof of Theorem 3.4 writes a conjugating element as a product of Coxeter generators, shows that consecutive permutations in the resulting path have descent numbers differing by at most 2 (Corollary 3.8), and then closes jumps of size 2 using cyclic rotations (Propositions 3.14 and 3.16). The paper also recalls background on Eulerian numbers and on known distribution results for statistics on conjugacy classes.
Significance. The claimed interval property is natural and attractive: it asserts that descent counts on each conjugacy class are gap-free, with a uniform minimum of 1, and it complements the distribution results of Gessel-Reutenauer, Fulman, and Kim-Lee with a simple structural statement. The constructive proof of Theorem 2.1 and the explicit recipe for hitting every intermediate value are concrete strengths. If the theorem is correct, the paper is a worthwhile contribution to the enumerative combinatorics of permutation statistics. However, the submitted proof of Theorem 3.4 relies on a false proposition (Proposition 3.5) and on mismatched left- versus right-multiplication statements (Proposition 3.6 and Theorem 3.7), so the central claim is not established as written. The argument appears repairable, but the repair requires substantial rewriting of Section 3.1.
major comments (3)
- [§3.1, Proposition 3.5] Proposition 3.5 is false as stated. In one-line notation, πs_i (right multiplication by the Coxeter generator) swaps the entries in positions i and i+1, whereas the hypothesis concerns the positions of the values i and i+1, which is the operation s_iπ. The proof itself says 'πs_i, π are identical except for the swap between the values i,i+1,' which describes s_iπ, not πs_i. A concrete counterexample is π=[2,1,4,3], i=2: here π^{-1}(2)=1 and π^{-1}(3)=4 are non-consecutive, des(π)=2, but πs_2=[2,4,1,3] has des(πs_2)=1. Since Proposition 3.5 is invoked in the proof of Theorem 3.7 (parts (A) and (B1)) to conclude des(π)=des(s_iπ), the proof of Theorem 3.7 is not valid as written.
- [§3.1, Proposition 3.6 and Theorem 3.7] Proposition 3.6 and Theorem 3.7 contain a left/right multiplication mismatch. Proposition 3.6 is stated for left multiplication, |des(π)-des(s_iπ)|≤1, but its proof and Table 1 verify the right-multiplication statement |des(π)-des(πs_i)|≤1: for instance, the pair 2143→2413 in the table is obtained by swapping the entries in positions 2 and 3. In the proof of Theorem 3.7(A), the sentence 'In 3.6 we prove |des(π)-des(πs_i)|≤1' is also not the needed estimate; the needed estimate is |des(s_iπ)-des(s_iπs_i)|≤1. Thus, even apart from the failure of Proposition 3.5, the manuscript does not contain a correct derivation of the claimed one-step bound |des(s_i^{-1}πs_i)-des(π)|≤1 in the cases covered by Theorem 3.7.
- [§3.1–§3.3, Corollaries 3.8–3.9 and proof of Theorem 3.4] Corollary 3.8, Corollary 3.9, and the final gap-closing step in the proof of Theorem 3.4 are unsupported as written. The local bound |des(π_{j+1})-des(π_j)|≤2 is the engine of the path argument, and the classification that a jump of 2 can occur only in case (a) with π^{-1}(i),π^{-1}(i+1) consecutive is used to justify the entire construction of σ_1. Because Proposition 3.5 is false and Theorem 3.7 is not proved, Corollary 3.8 is not established; its proof also asserts without proof that 'm_i is either added or removed' and that there is 'at most one change' from Des(s_iπ) to Des(s_iπs_i^{-1}). Consequently the two cases in the final proof of Theorem 3.4, which select σ_1 by conjugating σ_0 or σ_2 with powers of w, lack a valid foundation in the submitted text.
minor comments (5)
- [Section 1] The assertion 'for all n,k with 0≤k≤n, we have 0<A_{n,k}' is false because A_{n,n}=0; the maximum number of descents of a permutation in S_n is n-1, so the subsequent sentence should say 0≤k≤n-1.
- [Throughout] Several results state 1≤i≤n where the Coxeter generator s_i is defined only for 1≤i≤n-1 (Propositions 3.2, 3.5, 3.6, Theorem 3.7, Corollary 3.8, Proposition 3.16); this should be corrected throughout.
- [§3.3, Proposition 3.15] Proposition 3.15 is stated for 1≤t≤n, but for t=n the permutation σ=[1,2,...,n] has des(σ)=0, so the claimed des(σ)=1 fails; the statement should be restricted to 1≤t<n, which is the range used in the proof of Theorem 3.4.
- [Proof of Theorem 2.1] The phrase 'the Young diagram has at least two columns, i.e., k<n' is confusing: k is the number of rows, and the condition k<n follows from non-identity cycle type, not from a statement about the number of columns; this should be reworded.
- [Example 3.3(a)] The displayed conjugated permutation appears corrupted in the text ('[3 1 6 5 2 4]' is missing a digit), and should be typeset correctly.
Circularity Check
No circularity: the range theorem is a self-contained combinatorial derivation; the serious issue in the paper is a proof error, not circularity.
full rationale
The paper's central claim is a pure mathematical derivation from definitions. Theorem 2.1 constructs a permutation in every non-identity conjugacy class with exactly one descent, and Theorem 3.4 attempts to prove the interval property using one-step conjugation bounds, a local case analysis of the conjugation by a Coxeter generator, and cyclic-descent lemmas. No parameter is fitted, no numerical input is calibrated, and no external benchmark is used; the proof neither defines the target statistic in terms of the stated conclusion nor invokes a prior result of the authors as the load-bearing premise. There is a serious correctness issue in the proof as printed: Proposition 3.5 states a claim about pi s_i while the proof swaps the values i and i+1, which is the effect of s_i pi, and the example pi=[2,1,4,3], i=2 gives des(pi)=2 while des(pi s_i)=1, so the proposition is false as stated. This undermines the bound |des(pi_{j+1})-des(pi_j)|<=2 and the gap-closing argument, but a false statement is not an instance of circular reasoning: the conclusion is not equivalent by construction to an input, and no self-citation chain forces the result. Under the review rule that proof errors are correctness risks rather than circularity, the circularity score is 0.
Assumptions & free parameters
assumptions (3)
- standard math Conjugacy classes of S_n are exactly the sets of permutations with a fixed cycle type, i.e. a fixed partition of n.
- standard math The adjacent transpositions s_i=(i,i+1) generate S_n, and conjugation by any permutation preserves cycle type.
- ad hoc to paper For conjugation by a Coxeter generator, the descent number changes by at most 2, as stated in Corollary 3.8.
Cite this review
Pith. "Pith review of The range of the des statistic for conjugacy classes in $S_n$." pith.science (2026). https://pith.science/paper/722DU7KT
@misc{pith2026250522828,
author = {Pith},
title = {Pith review of: The range of the des statistic for conjugacy classes in $S_n$},
year = {2026},
howpublished = {\url{https://pith.science/paper/722DU7KT}},
note = {Machine review of arXiv:2505.22828}
}
abstract
We determine the range of the des statistic on every conjugacy class in the symmetric group $S_n$, prove that the minimum is $1$ (except for the identity class), and show that every intermediate value from $1$ to the maximum value is attained. We also demonstrate a constructive method to achieve every value in the range and discuss its combinatorial implications.
Reference graph
Works this paper leans on
-
[1]
Bona,Combinatorics of permutations, Chapman&Hall, CRC, 2004
M. Bona,Combinatorics of permutations, Chapman&Hall, CRC, 2004. 1
work page 2004
-
[2]
F. Brenti,Permutation enumeration, symmetric functions and unimodality, Pacific Journal of Mathematics Vol. 157 (1993), No.1. 1-28. 2
work page 1993
-
[3]
Cellini,Cyclic eulerian elements, European Journal of Combinatorics 19 (1998), 545-552
P. Cellini,Cyclic eulerian elements, European Journal of Combinatorics 19 (1998), 545-552. 2
work page 1998
-
[4]
M. Crochemore, J. Désarménien, & D. Perrin,A note on the Burrows-Wheeler transformation, Theoretical Computer Science 332 (2004), 567-572. 3
work page 2004
-
[5]
L. Euler,Methodus universalis series summandi ulterius promota, Euler Archive – All Works, 55 (1741), 147-158. 1
-
[6]
J. Fulman,The Distribution of Descents in Fixed Conjugacy Classes of the Symmetric Groups, Journal of Combinatorial Theory, Series A, Volume 84, Issue 2, 1998, Pages 171-180 2
work page 1998
-
[7]
I.M.GesselandC.Reutenauer, Counting Permutations with given cycle structure and descent set, J.Com- binatorial Theory (Ser. A) 64 (1993), 189-215. 2
work page 1993
-
[8]
G. B. Kim and S. Lee,Central limit theorem for descents in conjugacy classes of Sn , J. Combinatorial Theory (Ser. A), 169:105123, 2020. 2
work page 2020
Show all 11 references
- [9]
-
[10]
T. K. Petersen,Eulerian numbers, Birkhäuser Advanced Texts, Birkhäuser, 2015. 2
2015
-
[11]
T. J. Stieltjes,Sur la reduction en fraction continue d’une série procedant suivant les puissances descen- tantes d’une variable, Ann. Fac. Sc. Toulouse 3 (1889), 1-17. 1
Reviewed August 7, 2026 · model on record in the stance chip above.
Discussion (0). Sign in to comment.