Free self-decomposability and unimodality of the Fuss-Catalan distributions

Noriyoshi Sakuma; Wojciech M{\l}otkowski; Yuki Ueda

REVIEW 2 major objections 3 minor 24 references

Free self-decomposability and unimodality of the Fuss-Catalan distributions

T0 review · 2 major / 3 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper classifies which Fuss-Catalan distributions are freely self-decomposable: precisely the diagonal laws μ(p,p) with 1≤p≤2.

desk verdict The main classification is likely right and worth publishing, but the written proof has a genuine gap at p=2r and a false inequality in the p>2 unimodality argument. read the letter →

arxiv 1908.07887 v1 pith:OR2B5PSL submitted 2019-08-21 math.PR math.CO

Wojciech M{l}otkowski , Noriyoshi Sakuma , Yuki Ueda This is my paper · ORCID

classification math.PRmath.CO MSC 46L5460E07

keywords Fuss-Catalandistributionsfreeself-decomposabilityinfinitedivisibilityLévymeasurecumulantsunimodalityregularityL1class

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper settles which of the two-parameter Fuss-Catalan distributions μ(p,r), p≥1, 0

What carries the argument

The central object is the free Lévy density k_{p,r}(x)=xW_{p-r,r}(x), defined on the freely infinitely divisible region 0<r≤min{p/2,p-1}. It carries the argument because free self-decomposability is equivalent to k_{p,r}(x)dx being unimodal with mode 0; the sign of k'_{p,r}(x) just to the right of 0 is therefore decisive. Two auxiliary instruments do the heavy lifting: the parametrization of the support by ρ(φ) from Proposition 1.1 together with the asymptotic k_{p,r}(x)∼(1/π)sin(rπ/(p-r))$x^{{r/(p-r)}}$ as x→0+, which excludes every off-diagonal case in the proof; and, for μ(p,p), the integral representation of (n+2) binom(p,n+2) that makes the cumulant sequence {n r_n} manifestly conditionally positive definite for 1≤p≤2.

What would settle it

Run the cumulant-criterion test on the untouched p=2r edge: for μ(4,2) the free cumulants are A_n(2,2), and the 2×2 Hankel minor of {n r_n} equals (2·5)(4·42)−(3·14)^2 = 10·168−42² = −84<0, so Proposition 2.2 rules out free self-decomposability for this case; any off-diagonal pair whose Hankel minors are all nonnegative would refute the classification.

Watch

Extended reading notes

Core claim

The paper's central claim is a complete classification: for p≥1 and 0<r≤p, μ(p,r) is freely self-decomposable if and only if p=r and 1≤p≤2. The mechanism is the free Lévy measure. Every freely infinitely divisible Fuss-Catalan law has Lévy density k_{p,r}(x)=xW_{p-r,r}(x), and free self-decomposability is equivalent to k_{p,r}(x)dx being unimodal with mode 0. In the off-diagonal freely infinitely divisible region, the near-zero asymptotic of W_{p-r,r} forces k_{p,r} to increase from 0, so the mode-0 condition fails; on the diagonal μ(p,p), 1≤p≤2, the free cumulants are binomial coefficients and the sequence {n binom(p,n)} is conditionally positive definite, which by the cumulant criterion proves free self-decomposability. The paper also completes the description of free infinite divisibility for the family and gives the associated free Lévy-Khintchine representations.

Load-bearing premise

The proof of the off-diagonal exclusion depends on the asymptotic k_{p,r}(x)∼(1/π)sin(rπ/(p-r))$x^{{r/(p-r)}}$ near x=0; when p=2r the sine factor is 0, so the written argument supplies no sign information and leaves cases such as μ(4,2) uncovered.

Editorial extensions

If this is right

In the whole two-parameter family, free self-decomposability occurs only on the diagonal segment p=r with 1≤p≤2; every law with 0<r<p, including the freely infinitely divisible off-diagonal laws, fails it.
For 1≤p≤2 each μ(p,p) is not only freely self-decomposable but lies in the free L1 class: the remainder ρ_c in the dilation decomposition is itself freely self-decomposable for every c∈(0,1).
Free regularity holds exactly for 0<r≤min{p/2,p-1} and for p=r=1 or 2; in particular the diagonal interior 1<p<2 is freely self-decomposable yet not free regular.
The diagonal law μ(p,p) is unimodal for every p≥1, and the paper proves unimodality for μ(p,p-1), μ(2r,r), and for μ(2,r) with 1<r<r0; μ(1,r) is not unimodal for 0<r<1.
The paper's Conjecture 4.10 proposes a phase transition for every p>1: μ(p,r) is unimodal for r=p or 0<r≤r0(p), with r0(2)≈1.6756 supported by numerics.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The unhandled p=2r edge in the written proof can be settled directly: for μ(4,2), the cumulant Hankel test gives a negative 2×2 minor, so the classification survives that case even though the asymptotic argument does not cover it.
The proof strategy—reading free self-decomposability off the sign of the Lévy density near 0—suggests that for p=2r one needs the second term of the expansion of W_{p-r,r}; computing it would close the only written gap and could be checked against the same cumulant minors.
If the phase-transition conjecture holds, the curve r0(p) would split the off-diagonal region into a unimodal side near r=0 and a non-unimodal band below r=p, giving a natural upper boundary for any future extension of the free-self-decomposability classification.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Bottom line: the main classification (free self-decomposability iff 1 ≤ p = r ≤ 2) is very likely true, and the paper does a real service by correcting the earlier free-infinite-divisibility statement from [16]. But the written proof of the only-if direction has a genuine missing case at p=2r, and the unimodality claim for p>2 is argued with a false inequality. I would still send this to a referee; it deserves one.

What's new and good: the positive direction for μ(p,p), 1≤p≤2, is clean. The integral representations in Props 3.1 and 3.3 are straightforward, the free cumulant criterion is applied properly, and Theorem 3.9 actually puts μ(p,p) in free L1. The Lévy–Khintchine representation in Prop 3.6 is explicit and useful. The paper is also honest about correcting [16, Cor 7.1]; self-citation here is legitimate because the earlier statement was wrong.

Where it gets soft: Theorem 4.3 cites the small-x asymptotic k_{p,r}(x) ~ (1/π) sin(rπ/(p−r)) x^{r/(p−r)} from [9]. On the range 01, e.g. μ(4,2), μ(6,3)—and the asymptotic carries no sign information there. The paper doesn't treat it separately, so the written proof of the classification has a hole. I don't think the theorem is false; independent checks point the right way. But it needs an additional argument.

Also, Proposition 3.10 claims sin(pφ) ≥ sin(φ) on (0,π/p) for p>2; that is false near the right endpoint, where sin(pφ) is negative. So the unimodality proof for μ(p,p), p>2, rests on a false inequality. Maybe the conclusion can be saved by a more careful derivative analysis, but the present text doesn't do it. Proposition 4.6 for μ(2r,r) also has a range issue: (2r−1)φ is not confined to (0,π/2) for φ near π/(2r), so the sign analysis there is not fully justified.

Audience: free probabilists and anyone working with Fuss-Catalan/Raney distributions. The main structural theorem is the kind of thing that will be cited. If I were handling it, I would send it to a good referee in free probability, tell them the p=2r case and the unimodality proofs are the areas to scrutinize, and ask for revisions before acceptance.

Referee Report

2 major / 3 minor

Summary. The paper studies the two-parameter Fuss-Catalan (Raney) distributions μ(p,r), whose moments are the Raney numbers A_k(p,r). Its main results are a classification of free infinite divisibility (Theorem 1.3(1)), a classification of free self-decomposability (Theorem 1.3(2)), a classification of free regularity (Theorem 1.3(3)), and a number of unimodality statements for special families μ(p,p), μ(p,p−1), μ(2r,r), μ(1,r), and μ(2,r). The positive directions are based on integral representations of binomial-type sequences (Propositions 3.1 and 3.3) and on the free-cumulant criterion of Hasebe–Thorbjørnsen–Sakuma; the negative directions use the Forrester–Liu density asymptotics and determinantal positivity checks.

Significance. If the main classification is correct, Theorem 1.3(2) gives a complete characterization of free self-decomposability in the Fuss-Catalan family and corrects the earlier claim in [16] about free infinite divisibility. The explicit free Lévy–Khintchine representations (Propositions 3.6, 4.2, Corollary 4.8) and the hypergeometric computations are careful and useful. The arguments are mostly analytic and do not rely on numerical fitting, and the paper is honest about correcting an earlier oversight. However, the only-if direction of the central Theorem 4.3 has a genuine missing boundary case, so the main classification is not fully proved as written.

major comments (2)

[Section 4.3, Theorem 4.3] The only-if direction for 0<r≤min{p/2,p−1} uses the asymptotic k_{p,r}(x)∼(1/π)sin(rπ/(p−r))x^{r/(p−r)} as x→0+ and concludes that k'_{p,r}(x)≥0 near 0. On the subfamily p=2r, the sine factor is sin(π)=0, so the displayed leading term carries no sign information. This subfamily is nonempty in the range to be excluded (for instance μ(4,2), μ(6,3), and more generally μ(2r,r) for r>1); for r=1,p=2, there is the additional issue that W_{1,1} is the Dirac mass at 1 rather than a density, so formula (16) is not directly applicable. No separate analysis of p=2r appears in the manuscript, and therefore the classification of free self-decomposability in Theorem 1.3(2) is not established for this one-parameter family. The conclusion may well be true, but the written proof has a genuine gap.
[Section 4.5.2, Proposition 4.6] The proof asserts that rϕ and (2r−1)ϕ both lie in (0,π/2) for all ϕ∈(0,π/(2r)) when r>1. The first is true, but the second is false because (2r−1)π/(2r)>π/2 for r>1. Consequently, the term containing cos((2r−1)ϕ) is not necessarily nonnegative on the whole interval, and the inequality that drops this term is not justified. This invalidates the proof that g'_r(ϕ)≥0. The unimodality claim for μ(2r,r) needs a corrected argument; this does not affect the central free-self-decomposability classification, but it is a technical error in a stated result.

minor comments (3)

[Section 4.1, Theorem 4.1 proof] The case statement 'p>2 and p−1<r<p/2' is empty because p−1>p/2 for p>2; from the surrounding argument the intended case appears to be p/2<r<p−1 (with boundary cases handled separately). Please correct this typo, as it makes the exclusion of the region r≤p−1, r>p/2 unclear.
[Section 4.2, Proposition 4.2] The support formula (p−r)^{p−r}(p−r−1)^{1−(p−r)} gives the indeterminate form 0^0 when p−r=1, i.e. for μ(2,1); the boundary case p−r=1 should be excluded or treated separately.
[Section 3.3, equation (12) and Remark 3.7] The notation writes the Lebesgue measure as 'dx' inside the integrand after the indicator function, which is slightly misleading for a Lévy measure of the form k(x)dx/|x|; this is a presentational issue only.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the free self-decomposability classification is derived from external criteria and prior published inputs; the p=2r gap in Theorem 4.3 is a correctness issue, not a circular reduction.

full rationale

The paper's derivation chain is self-contained against external benchmarks. The central classification (Theorem 1.3(2)) rests on (i) the free cumulant formula r_n(μ(p,r)) = A_n(p−r,r), quoted from the prior published paper [15] and used only as an input; (ii) the characterization of free self-decomposability via conditional positive definiteness of {n r_n} and via unimodality of the free Lévy density, quoted from [13] and [4]; and (iii) the explicit density formula of Forrester–Liu (Proposition 1.1) and the asymptotic k_{p,r}(x) ∼ (1/π) sin(rπ/(p−r)) x^{r/(p−r)} from [9]. None of these inputs is defined in terms of the target classification, and no parameter is fitted to data. The self-citations [15] and [16] are prior published results used as lemmas; [16] is explicitly corrected for a separate issue, which is not circular. The proof does contain a genuine missing case at p=2r in Theorem 4.3, where the sine factor in the [9] asymptotic vanishes and the statement 'Hence k'_{p,r}(x) ≥ 0' is unsupported; this is a correctness gap, not a circular reduction, because the asymptotic and the criterion come from outside the paper and the missing case is not used to define the conclusion. Accordingly, no circular step is present.

Assumptions & free parameters 0 free parameters · 8 assumptions · 0 invented entities

The central claim rests on standard special-function identities, moment-sequence positivity, and established free-probability criteria. No parameters are fitted, and no entities are invented. The main load-bearing external input is the asymptotic in [9, Corollary 2.5], which has a degenerate case not addressed in the paper.

assumptions (8)

standard math Euler's reflection formula Γ(p)Γ(1−p)=π/sin(pπ) for p∉Z
Used in the integral evaluations in Propositions 3.1 and 3.3.
standard math Integral representation of the Gauss hypergeometric function 2F1(a,b,c;z) (equation (10))
Used to evaluate the free Levy-Khintchine representation in Proposition 3.6 and Corollary 4.8.
standard math Positive definiteness of a sequence with a positive integral kernel implies the sequence is a moment sequence
Underlies the passage from the integral representations to positive definiteness in Propositions 3.1 and 3.3.
domain assumption Bercovici-Voiculescu analytic extension theorem (Theorem 2.1) characterizing free infinite divisibility
Justifies analytic continuation of the free cumulant transform to C- after the local computations.
domain assumption Criterion from [13, Prop 2.2]: compactly supported µ is freely self-decomposable iff {n r_n(µ)} is conditionally positive definite
Used to convert the positivity of binomial sequences into free self-decomposability in Theorems 3.4 and 4.3.
domain assumption Free Levy measure criterion (Barndorff-Nielsen-Thorbjørnsen [4]): free self-decomposability implies ν(dx)=k(x)/|x| dx with k(x)dx unimodal with mode 0
Used in Remark 3.7, Theorem 3.9, and the exclusion argument in Theorem 4.3.
domain assumption Asymptotic k_{p,r}(x) ∼ (1/π) sin(rπ/(p−r)) x^{r/(p−r)} as x→0+ (Forrester-Liu [9, Corollary 2.5])
Load-bearing in the only-if direction of Theorem 4.3; degenerate when p=2r.
domain assumption Free regularity criterion via R-transform (Theorem 2.4 from [3])
Used to decide free regularity in Theorem 3.8 and Theorem 4.4.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Free self-decomposability and unimodality of the Fuss-Catalan distributions." pith.science (2026). https://pith.science/paper/OR2B5PSL

@misc{pith2026190807887,
  author       = {Pith},
  title        = {Pith review of: Free self-decomposability and unimodality of the Fuss-Catalan distributions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OR2B5PSL}},
  note         = {Machine review of arXiv:1908.07887}
}

abstract

We study properties of the Fuss-Catalan distributions $\mu(p,r)$, $p\geq1$, $0<r\leq p$: free infinite divisibility, free self-decomposability, free regularity and unimodality. We show that the Fuss-Catalan distribution $\mu(p,r)$ is freely self-decomposable if and only if $1 \leq p=r \leq 2$.

Figures

Figures reproduced from arXiv: 1908.07887 by the authors.

**Figure 2.** Phase transition References [1] G. E. Andrews, R. Askey, R. Roy, Special Functions. Encyclopedia of Mathematics and its Applications, 71, Cambridge University Press, Cambridge, 1999. xvi+664 pp. [2] O. Arizmendi, T. Hasebe, Classical and free infinite divisibility for Boolean stable laws. Proc. Amer. Math. Soc., 142 (2014), no. 5, 1621-1632. [3] O. Arizmendi, T. Hasebe, N. Sakuma, On the law of Free Subordinators. A… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 24 canonical work pages

[16]

Młotkowski, K

W . Młotkowski, K. A. Penson, Probability distribution s with binomial moments. Inﬁn. Dimens. Anal. Quantum Probab. Relat. Top., 17 (2014), no. 2, 1450014, 32 pp

work page 2014
[9]

P . J. Forrester, D.-Z. Liu, P . Zinn-Justin, Equilibriumproblems for Raney densities.Nonlinearity, 28 (2015), no. 7, 2265-2277

work page 2015
[1]

G. E. Andrews, R. Askey, R. Roy, Special Functions. Encyclopedia of Mathematics and its Applications , 71, Cambridge University Press, Cambridge, 1999. xvi+664 pp

work page 1999
[2]

Arizmendi, T

O. Arizmendi, T. Hasebe, Classical and free inﬁnite divi sibility for Boolean stable laws. Proc. Amer. Math. Soc., 142 (2014), no. 5, 1621-1632

work page 2014
[3]

Arizmendi, T

O. Arizmendi, T. Hasebe, N. Sakuma, On the law of Free Subo rdinators. ALEA, Lat. Am. J. Probab. Math. Stat., 10 (2013), no. 1, 271–291. 15

work page 2013
[4]

O. E. Barndorﬀ-Nielsen, S. Thorbjørnsen, Classical and free inﬁnite divisibility and Lévy processes. Quantum independent increment processes. II , 33–159, Lecture Notes in Math. , 1866, Springer, Berlin, 2006

work page 2006
[5]

Belinschi, M

S.T. Belinschi, M. Bożejko; F. Lehner; R. Speicher, The n ormal distribution is ⊞-inﬁnitely divisible. Adv. Math., 226 (2011), no. 4, 3677-3698

work page 2011
[6]

Benaych-Georges, Taylor expansions of R-transforms : application to supports and moments

F. Benaych-Georges, Taylor expansions of R-transforms : application to supports and moments. Indiana Univ. Math. J., 55 (2006), no. 2, 465–481

work page 2006

Show all 24 references

[7]

Bercovici, D

H. Bercovici, D. Voiculescu, Free convolution of measur es with unbounded support. Indiana Univ. Math. J., 42 (1993), no. 3, 733–773

1993
[8]

P . J. Forrester, D.-Z. Liu, Raney distributions and rand om matrix theory. J. Stat. Phys., 158 (2015), no. 5, 1051–1082

2015
[10]

R. L. Graham, D. E. Knuth, O. Patashnik, Concrete Mathem atics. A foundation for computer science , Addison-Wesley, 1994. xiv+657 pp

1994
[11]

Hasebe, Free inﬁnite divisibility for beta distribu tions and related ones

T. Hasebe, Free inﬁnite divisibility for beta distribu tions and related ones. Electron. J. Probab., 19 (2014), No. 81, 1-33

2014
[12]

Hasebe, S

T. Hasebe, S. Thorbjørnsen, Unimodality of the freely s elfdecomposable probability laws. J. Theoret. Probab., 29 (2016), no. 3, 922–940

2016
[13]

Hasebe, S

T. Hasebe, S. Thorbjørnsen, N.Sakuma, The normal distr ibution is freely self-decomposable. Int. Math. Res. Not. IMRN 2019, no. 6, 1758–1787

2019
[14]

Liu and R

J.-G. Liu and R. Pego, On generating functions of Hausdo rﬀ moment sequences. Trans. Am. Math. Soc. , 368 (2016), no. 12, 8499–8518

2016
[15]

Młotkowski, Fuss-Catalan numbers in noncommutativ e probability

W . Młotkowski, Fuss-Catalan numbers in noncommutativ e probability. Doc. Math., 15 (2010), 939–955

2010
[17]

Młotkowski, K

W . Młotkowski, K. A. Penson, K. Życzkowski, Densities of the Raney distributions. Doc. Math., 18 (2013), 1573–1596

2013
[18]

A. Nica, R. Speicher, Lectures on the combinatorics of f ree probability. London Mathematical Society Lecture Note Series, 335, Cambridge University Press, Cambridge, 2006. xvi+417 pp

2006
[19]

F. W . J. Olver, D. W . Lozier, R. F. Boisvert, C. W . Clark, NIST Handbook of Mathematical Functions. U.S. Department of Commerce, National Institute of Standards an d Technology, Washington, DC; Cambridge University Press, Cambridge 2010. xvi+951 pp

2010
[20]

R. P . Stanley, Catalan numbers. Cambridge University P ress, New Y ork, 2015. viii+215 pp

2015
[21]

Penson, K

K. Penson, K. Życzkowski, Product of Ginibre matrices: Fuss-Catalan and Raney distributions. Phys. Rev. E, 83, 061118 (2011)

2011
[22]

Sato, Class L of multivariate distributions of its su bclasses

K. Sato, Class L of multivariate distributions of its su bclasses. J. Multivariate Anal. , 10 (1980), no. 2, 207-232

1980
[23]

Sato, Fractional integrals and extensions of selfde composability.Lévy matters I, 1–91, 197–198, Lecture Notes in Math., 2001, Lévy Matters, Springer, Berlin, (2010)

K. Sato, Fractional integrals and extensions of selfde composability.Lévy matters I, 1–91, 197–198, Lecture Notes in Math., 2001, Lévy Matters, Springer, Berlin, (2010)

2010
[24]

Y amazato, Unimodality of inﬁnitely divisible distr ibution functions of class L, Ann

M. Y amazato, Unimodality of inﬁnitely divisible distr ibution functions of class L, Ann. Probab., 6 (1978), no. 4, 523–531. 16

1978

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [16]

Młotkowski, K

W . Młotkowski, K. A. Penson, Probability distribution s with binomial moments. Inﬁn. Dimens. Anal. Quantum Probab. Relat. Top., 17 (2014), no. 2, 1450014, 32 pp

work page 2014

[2] [9]

P . J. Forrester, D.-Z. Liu, P . Zinn-Justin, Equilibriumproblems for Raney densities.Nonlinearity, 28 (2015), no. 7, 2265-2277

work page 2015

[3] [1]

G. E. Andrews, R. Askey, R. Roy, Special Functions. Encyclopedia of Mathematics and its Applications , 71, Cambridge University Press, Cambridge, 1999. xvi+664 pp

work page 1999

[4] [2]

Arizmendi, T

O. Arizmendi, T. Hasebe, Classical and free inﬁnite divi sibility for Boolean stable laws. Proc. Amer. Math. Soc., 142 (2014), no. 5, 1621-1632

work page 2014

[5] [3]

Arizmendi, T

O. Arizmendi, T. Hasebe, N. Sakuma, On the law of Free Subo rdinators. ALEA, Lat. Am. J. Probab. Math. Stat., 10 (2013), no. 1, 271–291. 15

work page 2013

[6] [4]

O. E. Barndorﬀ-Nielsen, S. Thorbjørnsen, Classical and free inﬁnite divisibility and Lévy processes. Quantum independent increment processes. II , 33–159, Lecture Notes in Math. , 1866, Springer, Berlin, 2006

work page 2006

[7] [5]

Belinschi, M

S.T. Belinschi, M. Bożejko; F. Lehner; R. Speicher, The n ormal distribution is ⊞-inﬁnitely divisible. Adv. Math., 226 (2011), no. 4, 3677-3698

work page 2011

[8] [6]

Benaych-Georges, Taylor expansions of R-transforms : application to supports and moments

F. Benaych-Georges, Taylor expansions of R-transforms : application to supports and moments. Indiana Univ. Math. J., 55 (2006), no. 2, 465–481

work page 2006