arxiv: 2605.09655 · v1 · submitted 2026-05-10 · 💻 cs.IT · math.CO· math.IT· math.PR

Recognition: 2 theorem links

· Lean Theorem

Geometry of R\'enyi Entropy on the Majorization Lattice

Anuj Kumar Yadav , Yanina Y. Shkel

Authors on Pith no claims yet

Pith reviewed 2026-05-12 04:01 UTC · model grok-4.3

classification 💻 cs.IT math.COmath.ITmath.PR

keywords Rényi entropymajorization latticesubadditivitysupermodularitycomonotone couplingprobability distributionsinformation theory

0 comments

The pith

Rényi entropy is subadditive on the majorization lattice for every order α and supermodular for α equal to 0 or at least 1.

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

The paper examines how Rényi entropy changes when probability distributions are compared using the majorization order, which ranks distributions by how spread out their probabilities are. This order turns the set of sorted distributions into a complete lattice with well-defined meet and join operations. The authors first establish a relation between the comonotone coupling and the independent coupling of any collection of marginals. From this relation they derive that Rényi entropy is subadditive on the lattice for all α in [0, ∞]. They then identify the regime where the entropy is also supermodular, namely when α equals 0 or lies in [1, ∞].

Core claim

For every order α ∈ [0, ∞], the Rényi entropy is subadditive on the majorization lattice. The entropy is further supermodular on the same lattice when α belongs to {0} ∪ [1, ∞]. Both properties follow from a fundamental relation between the comonotone coupling and the independent coupling associated with any collection of marginal distributions.

What carries the argument

The fundamental relation between the comonotone coupling and the independent coupling of a collection of marginal distributions.

If this is right

Subadditivity supplies an upper bound on the Rényi entropy of the join of any two distributions in terms of their individual entropies.
Supermodularity for α ≥ 1 gives a convexity-type inequality that relates the entropy of the meet to the entropies of the inputs.
The same properties hold for the limiting cases α = 0 and α = ∞, recovering known behaviors of min-entropy and max-entropy on the lattice.
The results extend the classical subadditivity of Shannon entropy (the α = 1 case) to the entire one-parameter family of Rényi entropies.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same coupling relation might be usable to prove lattice properties for other f-divergences or entropies beyond the Rényi family.
Because majorization appears in quantum information and resource theories, the subadditivity result could translate into bounds on entropy-like quantities for quantum states ordered by majorization.
Numerical checks on small supports would quickly confirm or refute the claimed inequalities for intermediate α values.

Load-bearing premise

The majorization partial order forms a complete lattice on ordered probability distributions and the comonotone coupling relates to the independent coupling in the specific way needed for the entropy inequalities.

What would settle it

Select two concrete ordered distributions, compute their meet and join under majorization, then verify numerically whether the Rényi entropy of the join is at most the sum of the entropies of the two distributions for a chosen α outside the claimed supermodular range.

Figures

Figures reproduced from arXiv: 2605.09655 by Anuj Kumar Yadav, Yanina Y. Shkel.

**Figure 1.** Figure 1: (Example 1) : Supermodular behavior of the Rényi entropy for order α ∈ (0, 1). 0 1 2 3 4 5 6 7 8 9 10 −0.2 0 0.2 0.4 0.6 0.8 1 1.2 ·10−2 α ∆α(p, q) ∆α(p, q) vs α ∆α(p, q) ∆1(p, q) [PITH_FULL_IMAGE:figures/full_fig_p009_1.png] view at source ↗

**Figure 2.** Figure 2: (Example 2) : Submodular behavior of the Rényi entropy for order α ∈ (0, 1). Example 1 - Pair with ∆α(p, q) > 0: Let p, q ∈ P3 as, p = (0.6, 0.2, 0.2) q = (0.45, 0.4, 0.15) Consequently, the glb and lub are as follows, p ∧ q = (0.45, 0.35, 0.2) p ∨ q = (0.6, 0.25, 0.15) [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗

read the original abstract

Majorization is a stochastic ordering relation that compares the relative diversity of probability distributions with numerous applications in econometrics, spectral theory, and ecology. It is well-known that the majorization partial order forms a complete lattice on the set of ordered probability distributions. In this work, we study the properties of R\'enyi entropy on the majorization lattice. We establish a fundamental relation between the comonotone coupling and the independent coupling associated with a collection of marginal distributions. Consequently, we show that, for every order $ \alpha \in [0,\infty] $, the R\'enyi entropy is subadditive on the majorization lattice. We further characterize the supermodular regime, showing that R\'enyi entropy is supermodular on the majorization lattice for $ \alpha \in \{0\} \cup [1,\infty] $.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper uses a comonotone-independent coupling to prove Rényi entropy is subadditive on the majorization lattice for every alpha, plus a clean supermodular characterization for alpha at 0 or >=1.

read the letter

The main thing to know is that Rényi entropy satisfies subadditivity with respect to the join in the majorization lattice across the entire range of alpha, and the authors isolate the supermodular regime to alpha equals zero or at least one. They start from the known fact that ordered probability vectors form a complete lattice under majorization, then introduce a relation between the comonotone coupling and the independent coupling of marginals. From that relation they derive the subadditivity inequality directly. The supermodular part follows by checking the lattice meet and join behavior in the indicated ranges. This is new in the sense that prior majorization work on entropies did not pin down the full subadditivity or the exact supermodular window for Rényi. It sits well because it reuses standard lattice properties without inventing new structure, and the coupling step is a natural way to compare joint and marginal behaviors. The soft spot is the uniformity of the coupling argument across alpha. Rényi entropy changes its convexity properties at alpha equals one, so the same coupling relation has to deliver the inequality both when the functional is concave and when it is not. The abstract states the result for all alpha but does not show the case split, so any gap in handling 0 less than alpha less than 1 would undercut the central claim. The lattice completeness part is not the issue; it is already established. This paper is for people who work with majorization in information theory, inequality measurement, or diversity indices and want to see how Rényi fits inside the lattice. A reader already comfortable with couplings and lattice theory will get the most out of it. It is worth sending to referees because the claims are concrete and the supporting machinery is standard; a careful check of the derivations will settle whether the coupling step holds everywhere it needs to.

Referee Report

2 major / 1 minor

Summary. The paper claims that the majorization partial order forms a complete lattice on ordered probability distributions, establishes a relation between comonotone and independent couplings of marginals, and uses this to prove that Rényi entropy H_α is subadditive on the lattice for every α ∈ [0, ∞]. It further shows that H_α is supermodular on the lattice precisely when α ∈ {0} ∪ [1, ∞].

Significance. If the coupling relation is shown to deliver the subadditivity inequality uniformly, the work would supply a lattice-geometric view of Rényi entropy that distinguishes its subadditive and supermodular regimes. This could connect majorization theory with information-theoretic functionals in a way that is useful for applications in diversity measurement and stochastic ordering.

major comments (2)

[Section establishing the coupling relation and the subadditivity theorem] The central subadditivity claim for all α ∈ [0, ∞] rests on the comonotone-independent coupling relation. The manuscript must supply the explicit steps showing that this relation implies H_α(x ∨ y) ≤ H_α(x) + H_α(y) when 0 < α < 1, where the functional form of H_α is neither convex nor concave in the standard sense used for α ≥ 1. Without this verification the uniform claim is not yet load-bearing.
[Section on supermodular regime] The supermodularity characterization is stated only for α ∈ {0} ∪ [1, ∞]. The paper should either prove that supermodularity fails for 0 < α < 1 (e.g., via an explicit counter-example on the lattice) or clarify why the coupling argument does not extend to supermodularity in that interval.

minor comments (1)

[Introduction / preliminaries] Notation for the join operation ∨ on the lattice and for the Rényi entropy functional should be introduced with a short reminder of the standard definitions before the main theorems.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive feedback on our manuscript. The comments highlight areas where additional detail will strengthen the presentation of the coupling relation and the characterization of supermodularity. We address each major comment below and will incorporate the suggested clarifications and examples in the revised version.

read point-by-point responses

Referee: [Section establishing the coupling relation and the subadditivity theorem] The central subadditivity claim for all α ∈ [0, ∞] rests on the comonotone-independent coupling relation. The manuscript must supply the explicit steps showing that this relation implies H_α(x ∨ y) ≤ H_α(x) + H_α(y) when 0 < α < 1, where the functional form of H_α is neither convex nor concave in the standard sense used for α ≥ 1. Without this verification the uniform claim is not yet load-bearing.

Authors: We agree that the derivation for 0 < α < 1 merits explicit expansion. The comonotone-independent coupling relation is established in general form in the manuscript and is used to identify the join x ∨ y with the comonotone coupling of the marginals. For 0 < α < 1 the Rényi functional is handled via the monotonicity of t ↦ t^{1/α} (which is decreasing in this range) together with the majorization ordering induced by the coupling. In the revision we will insert a dedicated lemma that isolates this case, writing out the chain of inequalities from the coupling probabilities to the final bound H_α(x ∨ y) ≤ H_α(x) + H_α(y). This will make the uniform claim fully self-contained. revision: yes
Referee: [Section on supermodular regime] The supermodularity characterization is stated only for α ∈ {0} ∪ [1, ∞]. The paper should either prove that supermodularity fails for 0 < α < 1 (e.g., via an explicit counter-example on the lattice) or clarify why the coupling argument does not extend to supermodularity in that interval.

Authors: The coupling argument supplies the join inequality uniformly but does not control the meet inequality needed for supermodularity when 0 < α < 1, because the functional is no longer Schur-concave in the same manner. We will therefore add an explicit counter-example in the revised manuscript. Using two three-dimensional ordered probability vectors whose join and meet are readily computed, we will exhibit numerical values for a fixed α ∈ (0,1) that violate H_α(x ∨ y) + H_α(x ∧ y) ≥ H_α(x) + H_α(y). This counter-example will be placed immediately after the supermodularity theorem to justify the precise regime {0} ∪ [1, ∞]. revision: yes

Circularity Check

0 steps flagged

No circularity; derivation uses derived coupling relation on standard lattice

full rationale

The paper first states the well-known completeness of the majorization lattice on ordered distributions, then establishes a relation between comonotone and independent couplings (a first-principles step), and only then concludes subadditivity of Rényi entropy for all α. No step reduces by definition or by self-citation to the target claim itself; the coupling relation is presented as newly derived rather than fitted or renamed from prior results of the same authors. The supermodularity characterization for specific α ranges follows similarly without circular reduction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The work relies on the well-known lattice structure of majorization and standard coupling properties in probability; no free parameters or invented entities are introduced.

axioms (2)

domain assumption Majorization partial order forms a complete lattice on ordered probability distributions
Explicitly stated as well-known in the abstract
standard math Existence of comonotone and independent couplings for marginal distributions
Used as the basis for establishing the fundamental relation

pith-pipeline@v0.9.0 · 5449 in / 1212 out tokens · 66326 ms · 2026-05-12T04:01:07.131799+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

We establish a fundamental relation between the comonotone coupling and the independent coupling... Consequently, we show that, for every order α∈[0,∞], the Rényi entropy is subadditive on the majorization lattice.
IndisputableMonolith/Foundation/AlexanderDuality.lean alexander_duality_circle_linking unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Bapat showed that the majorization order forms a complete lattice... Cicalese and Vaccaro proved that Shannon entropy is supermodular and subadditive on this lattice.

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

40 extracted references · 40 canonical work pages

[1]

Marshall, I

A. Marshall, I. Olkin, and B. Arnold,Inequalities: Theory of Majorization and Its Applications, ser. Springer Series in Statistics. Springer New York, 2010. [Online]. Available: https://books.google.ch/ books?id=jE11p6ziLJ0C

work page 2010
[2]

Some simple inequalities satisfied by convex functions,

G. H. Hardy, “Some simple inequalities satisfied by convex functions,” Messenger Math., vol. 58, pp. 145–152, 1929

work page 1929
[3]

Hardy, J

G. Hardy, J. Littlewood, and G. Pólya,Inequalities, ser. Cambridge Mathematical Library. Cambridge University Press, 1952. [Online]. Available: https://books.google.ch/books?id=t1RCSP8YKt8C

work page 1952
[4]

Über eine klasse von mittelbildungen mit anwendungen auf die determinantentheorie,

I. Schur, “Über eine klasse von mittelbildungen mit anwendungen auf die determinantentheorie,”Sitzungsberichte der Berliner Mathematischen Gesellschaft, vol. 22, pp. 9–20, 1923

work page 1923
[5]

Methods of measuring the concentration of wealth,

M. O. Lorenz, “Methods of measuring the concentration of wealth,” Publications of the American Statistical Association, vol. 9, no. 70, pp. 209–219, 1905. [Online]. Available: http://www.jstor.org/stable/2276207

work page arXiv 1905
[6]

Pigou,Wealth and Welfare

A. Pigou,Wealth and Welfare. Creative Media Partners, LLC,

work page
[7]

Available: https://books.google.co.in/books?id=1qM_ swEACAAJ

[Online]. Available: https://books.google.co.in/books?id=1qM_ swEACAAJ

work page
[8]

The measurement of the inequality of incomes,

H. Dalton, “The measurement of the inequality of incomes,”The Economic Journal, vol. 30, no. 119, pp. 348–361, 1920. [Online]. Available: http://www.jstor.org/stable/2223525

work page arXiv 1920
[9]

Rényi divergence and majorization,

T. van Erven and P. Harremoës, “Rényi divergence and majorization,” in2010 IEEE International Symposium on Information Theory, 2010, pp. 1335–1339

work page 2010
[10]

A finite sufficient set of conditions for catalytic majorization,

D. Elkouss, A. G. Maity, A. Nema, and S. Strelchuk, “A finite sufficient set of conditions for catalytic majorization,”Communications Physics,

work page
[11]

Available: https://doi.org/10.1038/s42005-026-02583-x

[Online]. Available: https://doi.org/10.1038/s42005-026-02583-x

work page doi:10.1038/s42005-026-02583-x
[12]

Matrix majorization in large samples,

M. U. Farooq, T. Fritz, E. Haapasalo, and M. Tomamichel, “Matrix majorization in large samples,”IEEE Transactions on Information Theory, vol. 70, no. 5, pp. 3118–3144, 2024

work page 2024
[13]

Mimo transceiver design via majorization theory,

D. P. Palomar and Y . Jiang, “Mimo transceiver design via majorization theory,”Foundations and Trends in Communications and Information Theory, vol. 3, no. 4-5, pp. 331–551, 06 2007. [Online]. Available: https://doi.org/10.1561/0100000018

work page doi:10.1561/0100000018 2007
[14]

Conditions for a class of entanglement transformations,

M. A. Nielsen, “Conditions for a class of entanglement transformations,” Phys. Rev. Lett., vol. 83, pp. 436–439, Jul 1999. [Online]. Available: https://link.aps.org/doi/10.1103/PhysRevLett.83.436

work page doi:10.1103/physrevlett.83.436 1999
[15]

Quantum majorization and a complete set of entropic conditions for quantum thermodynamics,

G. Gour, D. Jennings, F. Buscemi, R. Duan, and I. Marvian, “Quantum majorization and a complete set of entropic conditions for quantum thermodynamics,”Nature Communications, vol. 9, no. 1, p. 5352,

work page
[16]

Quantum majoriza- tion and a complete set of entropic conditions for quantum thermodynamics

[Online]. Available: https://doi.org/10.1038/s41467-018-06261-7

work page doi:10.1038/s41467-018-06261-7
[17]

Inequality measurement with grouped data: Parametric and non-parametric methods,

V . Jorda, J. M. Sarabia, and M. Jäntti, “Inequality measurement with grouped data: Parametric and non-parametric methods,”Journal of the Royal Statistical Society Series A: Statistics in Society, vol. 184, no. 3, pp. 964–984, 07 2021. [Online]. Available: https://doi.org/10.1111/rssa.12702

work page doi:10.1111/rssa.12702 2021
[18]

Multidimensional inequality measurement via optimal transport,

Y . Fan, M. Henry, B. Pass, and J. A. Rivero, “Multidimensional inequality measurement via optimal transport,”The Review of Economics and Statistics, pp. 1–45, 11 2024. [Online]. Available: https://doi.org/10.1162/rest_a_01539

work page doi:10.1162/rest_a_01539 2024
[19]

Newton– simpson-type inequalities via majorization,

S. I. Butt, I. Javed, P. Agarwal, and J. J. Nieto, “Newton– simpson-type inequalities via majorization,”Journal of Inequalities and Applications, vol. 2023, no. 1, p. 16, 2023. [Online]. Available: https://doi.org/10.1186/s13660-023-02918-0

work page doi:10.1186/s13660-023-02918-0 2023
[20]

Minimum-entropy couplings and their applications,

F. Cicalese, L. Gargano, and U. Vaccaro, “Minimum-entropy couplings and their applications,”IEEE Transactions on Information Theory, vol. 65, no. 6, pp. 3436–3451, 2019

work page 2019
[21]

Efficient approximate minimum entropy coupling of multiple probability distributions,

C. T. Li, “Efficient approximate minimum entropy coupling of multiple probability distributions,”IEEE Transactions on Information Theory, vol. 67, no. 8, pp. 5259–5268, 2021

work page 2021
[22]

A tighter approximation guarantee for greedy minimum entropy coupling,

S. Compton, “A tighter approximation guarantee for greedy minimum entropy coupling,” in2022 IEEE International Symposium on Informa- tion Theory (ISIT), 2022, pp. 168–173

work page 2022
[23]

Information spectrum converse for minimum entropy couplings and functional representations,

Y . Y . Shkel and A. Kumar Yadav, “Information spectrum converse for minimum entropy couplings and functional representations,” in2023 IEEE International Symposium on Information Theory (ISIT), 2023, pp. 66–71

work page 2023
[24]

Infinite divisibility of information,

C. T. Li, “Infinite divisibility of information,”IEEE Transactions on Information Theory, vol. 68, no. 7, pp. 4257–4271, 2022

work page 2022
[25]

Minimum-entropy coupling approximation guarantees beyond the majorization barrier,

S. Compton, D. Katz, B. Qi, K. Greenewald, and M. Kocaoglu, “Minimum-entropy coupling approximation guarantees beyond the majorization barrier,” inProceedings of The 26th International Conference on Artificial Intelligence and Statistics, ser. Proceedings of Machine Learning Research, F. Ruiz, J. Dy, and J.-W. van de Meent, Eds., vol. 206. PMLR, 25–27 Apr...

work page 2023
[26]

Completely positive linear maps on complex matrices,

R. Bapat, “Majorization and singular values. iii,”Linear Algebra and its Applications, vol. 145, pp. 59–70, 1991. [Online]. Available: https://www.sciencedirect.com/science/article/pii/0024379591902877

work page arXiv 1991
[27]

Submodular functions and convexity,

L. LovÃ¡sz, “Submodular functions and convexity,” pp. 235–257, March 1983. [Online]. Available: https://ideas.repec.org/h/spr/sprchp/ 978-3-642-68874-4_10.html

work page 1983
[28]

Fujishige,Submodular Functions and Optimization, ser

S. Fujishige,Submodular Functions and Optimization, ser. Annals of Discrete Mathematics. North Holland, 1991. [Online]. Available: https://books.google.ch/books?id=a5PEEQPLphYC

work page 1991
[29]

Polymatroidal dependence structure of a set of random variables,

——, “Polymatroidal dependence structure of a set of random variables,”Information and Control, vol. 39, no. 1, pp. 55–72, 1978. [Online]. Available: https://www.sciencedirect.com/science/article/pii/ S001999587891063X

work page 1978
[30]

A new outlook on shannon’s information measures,

R. Yeung, “A new outlook on shannon’s information measures,”IEEE Transactions on Information Theory, vol. 37, no. 3, pp. 466–474, 1991

work page 1991
[31]

The size of a share must be large,

L. Csirmaz, “The size of a share must be large,”Journal of Cryptology, vol. 10, no. 4, pp. 223–231, 1997. [Online]. Available: https://doi.org/10.1007/s001459900029

work page doi:10.1007/s001459900029 1997
[32]

Jukna,Extremal Combinatorics: With Applications in Computer Science, 2nd ed

S. Jukna,Extremal Combinatorics: With Applications in Computer Science, 2nd ed. Springer Publishing Company, Incorporated, 2011

work page 2011
[33]

Supermodularity and subadditivity prop- erties of the entropy on the majorization lattice,

F. Cicalese and U. Vaccaro, “Supermodularity and subadditivity prop- erties of the entropy on the majorization lattice,”IEEE Transactions on Information Theory, vol. 48, no. 4, pp. 933–938, 2002

work page 2002
[34]

The construction of huffman codes is a submodular (

D. S. Parker and P. Ram, “The construction of huffman codes is a submodular ("convex") optimization problem over a lattice of binary trees,”SIAM Journal on Computing, vol. 28, no. 5, pp. 1875–1905,

work page 1905
[35]

Available: https://doi.org/10.1137/S0097539796311077

[Online]. Available: https://doi.org/10.1137/S0097539796311077

work page doi:10.1137/s0097539796311077
[36]

Extremal elements of a sublattice of the majorization lattice and approximate majorization,

C. Massri, G. Bellomo, F. Holik, and G. M. Bosyk, “Extremal elements of a sublattice of the majorization lattice and approximate majorization,”Journal of Physics A: Mathematical and Theoretical, vol. 53, no. 21, p. 215305, may 2020. [Online]. Available: https://doi.org/10.1088/1751-8121/ab8674

work page doi:10.1088/1751-8121/ab8674 2020
[37]

The geometry of coalition power: Majorization, lattices, and displacement in multiwinner elections,

Q. Guo, Y . Hu, and R. Zhang, “The geometry of coalition power: Majorization, lattices, and displacement in multiwinner elections,”arXiv preprint arXiv:2601.16723, 2026

work page arXiv 2026
[38]

An information theoretic approach to probability mass function truncation,

F. Cicalese, L. Gargano, and U. Vaccaro, “An information theoretic approach to probability mass function truncation,” in2019 IEEE Inter- national Symposium on Information Theory (ISIT), 2019, pp. 702–706

work page 2019
[39]

Information theoretic measures of distances and their economet- ric applications,

——, “Information theoretic measures of distances and their economet- ric applications,” in2013 IEEE International Symposium on Information Theory, 2013, pp. 409–413

work page 2013
[40]

Economics and information theory,

H. Theil, “Economics and information theory,” 1967. APPENDIX A. Proof of Lemma 1 Proof.Givenp,q∈ P n. LetUbe a uniformly distributed continuous random variable on[0,1]i.e.,U∼Unif([0,1]). Now, define random variablesXandYon the support set {1,2, . . . , n}as X≜min   i: iX j=1 pj ≥U    Y≜min   i: iX j=1 qj ≥U    Thus, we have thatX∼p,Y∼qand they...

work page 1967