REVIEW 3 major objections 3 minor 12 references
Union-closed families minimize k-chains by taking sets as large as possible.
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 →
Union-closed supersaturation: fixed-size union-closed families minimize k-chains exactly when they are top-aligned, with uniqueness for m>n and positive minimum.
T0 review reviewed 2026-08-01 challenge →
load-bearing objection A genuine new supersaturation theorem for union-closed families; the proof is largely sound and the alleged circularity is not actually there. the 3 major comments →
Supersaturation in union-closed families of sets
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
Core claim
For positive integers k and n with 2≤k≤n+1, let c(k,m,n) be the minimum number of k-chains in a union-closed family with universe [n] and m members. The paper's main theorem asserts that c(k,m,n) is attained by a top-aligned family — one occupying the highest layers of the Boolean lattice, so that every member set is at least as large as every absent set. It further claims that for m>n with c(k,m,n)>0, this minimizer is unique. The argument routes an arbitrary family through iterated up-compression (which preserves union-closure and does not increase chain counts) and then through an iterative swap that replaces the smallest present set by a largest absent set; the swap is shown to preserve
What carries the argument
The proof rests on two operators. T1 is the composition of 'up-compressions' u_1,...,u_n, each of which adds a fixed element x to a set unless that larger set is already present; iterated compression preserves union-closure, preserves universe and size, and never increases the number of k-chains, eventually producing an upward-closed family. T2 is an iterative swap V: while the family is not top-aligned, remove the lexicographically first smallest member α_A and insert the lexicographically first largest missing member β_A. Lemma 2.2.1 shows V preserves upward-closedness; Lemma 2.2.2 bounds the chain-count change by an injection from chains through β_A to chains through α_A, with strict decr
Load-bearing premise
The strict-decrease and uniqueness proof depends on the claim that if c(k,m,n)>0 then k≤n−|α_A|+1 for the smallest set in the family; as written, that claim is justified by the top-aligned minimization that the proof is meant to establish, so the uniqueness result rests on a circular step unless a noncircular argument is supplied.
What would settle it
For n=5, m=8, k=3, exhaustively enumerate all union-closed families of 8 subsets of [5]. The theorem predicts the top-aligned family (all sets of size at least 4 plus two 3-sets) uniquely minimizes the number of 3-chains. A non-top-aligned family with the same chain count would refute the uniqueness half; if none exists, the only failure point is the circular justification identified above.
If this is right
- The exact minimum c(k,m,n) is determined for all allowed k and m once the top-aligned family is constructed, giving a union-closed analogue of classical supersaturation results.
- For m>n and c(k,m,n)>0, any union-closed family attaining the minimum must be top-aligned, so the extremal family can be described explicitly as the highest m layers of the power set.
- In the separating case, where m≥n automatically, the same minimum and uniqueness hold; for m=n, a separating family must be top-aligned to minimize 2-chains.
- Since the extremal family is top-aligned, its chain count can be computed by counting chains among the top layers of the Boolean lattice.
- The two-operator framework suggests a general strategy for supersaturation problems: push families upward without increasing the statistic, then remove a smallest set and add a largest missing set.
Where Pith is reading between the lines
- The uniqueness claim is more brittle than the minimization claim: it requires a noncircular proof that a positive minimum forces k≤n−|α_A|+1 for the smallest set in the family, for example via a binomial bound on chain-free families.
- The same two-operator scheme may extend to weighted or to multi-chain counts, since the injection argument in Lemma 2.2.2 is not tied to unweighted chains.
- The corollary suggests that 'separating' adds no new obstructions to extremality; one could test whether weaker separation assumptions alone preserve uniqueness for m>n.
- The theorem leaves the zero-minimum regime open; a natural extension would determine whether uniqueness can fail there.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies union-closed families of subsets of [n] and claims an exact supersaturation theorem: among all union-closed families with universe [n] and size m, the number of k-chains is minimized by top-aligned families, and when the minimum is positive and m>n these are the only minimizers. The proof combines an iterated up-compression operator T1 with a second operator V that repeatedly replaces a lexicographically first smallest member by a lexicographically first largest missing set. The authors show that T1 does not increase the number of k-chains, V preserves upward-closure and also does not increase the number of k-chains, and that the resulting top-aligned family gives the claimed minimum. A separate section treats uniqueness and a corollary for separating families.
Significance. If the main theorem is correct, it gives the exact extremal value and a complete extremal characterization for chain supersaturation in union-closed families, complementing Erdős's bound and the Kleitman–Samotij theory in the union-closed setting. The compression strategy is coherent and the small cases I checked are consistent with the statement. No machine-checked proofs or code are provided, but the arguments are elementary and largely self-contained. The proof has several compressed steps that need expansion, but I found no direct counterexample and the flagged circularity in Lemma 2.2.2 is, in my reading, not actually circular. The main mathematical idea is sound and publishable after a careful revision.
major comments (3)
- [Lemma 2.2.2, strictness step] The parenthetical justification for k≤n−|α_A|+1 is too terse and can be misread as circular. It is repairable by a non-circular argument: from m≤Σ_{i=0}^{n−|α_A|} C(n,i), any top-aligned family of size m has minimum set size at least |α_A|; otherwise it would contain all sets of size ≥|α_A| plus at least one smaller set, giving more than Σ_{i=|α_A|}^n C(n,i)=Σ_{i=0}^{n−|α_A|} C(n,i) ≥ m sets. Hence its longest chain has length at most n−|α_A|+1. So if k>n−|α_A|+1, every top-aligned family of size m has zero k-chains, contradicting c(k,m,n)>0. Please replace the parenthetical with this explicit derivation.
- [End of §2.2 / statement of Theorem] The proof at the end of §2.2 establishes that for every union-closed A0 there exists a top-aligned family A_min with |C(A_min,k)|≤|C(A0,k)|. This gives the existence of a top-aligned minimizer. However, the theorem's first sentence asserts that the minimum is attained 'whenever the family is top-aligned.' To justify this one must prove that all top-aligned families of the same size m have the same number of k-chains. This is true (any chain contains at most one set from the partially selected boundary layer, and the contribution of a boundary set depends only on its size), but it is not stated or proved. The uniqueness argument in §2.3 also implicitly uses equality |C(T1(A0),k)|=c(k,m,n) for the particular top-aligned family T1(A0), which again requires this comparison among top-aligned families. Add a short lemma.
- [§2.3, m>n+1 case] Several load-bearing assertions in the uniqueness proof are left unjustified. (i) The existence of Y with q≤|Y|≤q+1, r∉Y and Y∪{r}∉T1^-(A0) needs proof: T1^- has no sets of size below q, and a changed set of size at least q+2 would already lie in the top-aligned family T1. (ii) The existence of Z of size q+2 with r∉Z also needs proof; it follows because every such set must already be in T1^-, since its preimage under u_r cannot be the corresponding size-(q+3) superset containing r. (iii) The bound m≤Σ_{i=0}^{n−q−1} C(n,i) is in fact an equality and should be derived from the definition of q and the fact that T1 is top-aligned. These are routine but currently appear as unsupported leaps in the central uniqueness argument.
minor comments (3)
- [§2.3, definition of T1^-] The notation T1^- = u_{r-1}∘...∘u_0 appears to be a typo: the operators were defined as u_1,...,u_n. Either define u_0 as the identity and reindex, or write u_{r-1}∘...∘u_1.
- [§2.3, final paragraph] There is a missing parenthesis: '|C(T1(A0, k))|' should be '|C(T1(A0), k)|'.
- [§2.2, Figure 2.2.1] The figure is only a schematic; it would help to label the layers and the sets α_A and β_A explicitly in the caption.
Circularity Check
No significant circularity: Lemma 2.2.2's c(k,m,n)>0 step is independent of the theorem, and the main proof is self-contained.
full rationale
The main theorem is obtained by explicit, externally checkable transformations: T1 (iterated up-compression) with Lemma 2.1.1 and T2 (iterated V) with Lemmas 2.2.1-2.2.2. The only step that might appear circular is the parenthetical in Lemma 2.2.2: 'From this, it follows that k≤n−|α_A|+1. (Otherwise, no top-aligned family with such size and universe would have a k-chain, contradicting that c(k,m,n)>0.)' This is not circular. Positivity of c(k,m,n) means every union-closed family with universe [n] and size m contains a k-chain; top-aligned families of that size and universe are union-closed, so they too must contain one. If k>n−|α_A|+1, then any top-aligned family has smallest member of size at least |α_A| and hence longest chain at most n−|α_A|, so it would contain no k-chain, a contradiction. This uses only the definition of c and the elementary structure of top-aligned families, not the theorem being proved. The analogous k≤n−q step in §2.3 is valid for the same reason. The author's earlier upper bound [2] is cited for context but is not a premise of this derivation; Reimer's compression properties and Samotij's external theorem are used as standard supporting facts. No 'prediction' reduces by construction to a fitted parameter or to a self-citation chain, so the circularity score is 0.
Axiom & Free-Parameter Ledger
axioms (3)
- standard math Reimer's theorem: iterating up-compressions u_n∘...∘u_1 of a union-closed family preserves union-closedness and yields an upward-closed family.
- standard math Erdős's theorem on the maximum size of a family with no k-chain: the maximum is the sum of the k−1 largest binomial coefficients.
- domain assumption All top-aligned families of the same size m have the same number of k-chains.
Cite this review
Pith. "Pith review of Supersaturation in union-closed families of sets." pith.science (2026). https://pith.science/paper/N4AHWGXU
@misc{pith2026260726425,
author = {Pith},
title = {Pith review of: Supersaturation in union-closed families of sets},
year = {2026},
howpublished = {\url{https://pith.science/paper/N4AHWGXU}},
note = {Machine review of arXiv:2607.26425}
}
abstract
Let $k$ and $n$ be positive integers such that $2 \leq k \leq n+1$. We prove that the number of $k$-chains in a union-closed family with universe $[n]$ and size $m$ is minimized when its member sets are largest possible. We also show that, whenever the minimum is nonzero and $m>n$, there are no other minimizing families.
Figures
Reference graph
Works this paper leans on
-
[1]
Balogh and A.Z
J. Balogh and A.Z. Wagner,Kleitman ’s conjecture about families of given size minimizing the number ofk-chains, Adv. Math.330(2018), 229–252
2018
-
[2]
Bouchard,An upper bound for union-closed family size, preprint (2025), arXiv:2511.10608
C. Bouchard,An upper bound for union-closed family size, preprint (2025), arXiv:2511.10608
arXiv 2025
-
[3]
Bruhn and O
H. Bruhn and O. Schaudt,The journey of the union-closed sets conjecture, Graphs Combin.31 (2015), 2043–2074
2015
-
[4]
S. Das, W. Gan, and B. Sudakov,Sperner’s theorem and a problem of Erd˝ os, Katona and Kleitman, Combin. Probab. Comput.24(2015), 585–608
2015
-
[5]
Dove, J.R
A.P. Dove, J.R. Griggs, R.J. Kang, and J.-S. Sereni,Supersaturation in the Boolean lattice, Integers 14A(2014), A4
2014
-
[6]
Erd˝ os,On a lemma of Littlewood and Offord, Bull
P. Erd˝ os,On a lemma of Littlewood and Offord, Bull. Amer. Math. Soc.51(1945), 898–902
1945
-
[7]
Falgas-Ravry,Minimal weight in union-closed families, Electron
V. Falgas-Ravry,Minimal weight in union-closed families, Electron. J. Combin.18(1)(2011), P95
2011
-
[8]
J. Gilmer,A constant lower bound for the union-closed sets conjecture, preprint (2022), arXiv:2211.09055
Pith/arXiv arXiv 2022
-
[9]
Kleitman,A conjecture of Erd˝ os-Katona on commensurable pairs among subsets of an n-set (Theory of Graphs: Proc
D. Kleitman,A conjecture of Erd˝ os-Katona on commensurable pairs among subsets of an n-set (Theory of Graphs: Proc. Colloq. Tihany 1966), Academic Press (1968), 215–218
1966
-
[10]
Reimer,An average set size theorem, Combin
D. Reimer,An average set size theorem, Combin. Probab. Comput.12(2003), 89–93
2003
-
[11]
Samotij,Subsets of posets minimising the number of chains, Trans
W. Samotij,Subsets of posets minimising the number of chains, Trans. Amer. Math. Soc.371 (2019), 7259–7274
2019
-
[12]
Sperner,Ein Satz ¨ uber Untermengen einer endlichen Menge, Math
E. Sperner,Ein Satz ¨ uber Untermengen einer endlichen Menge, Math. Z.27(1928), 544–548. 6
1928
This paper was first reviewed by deepseek-v4-flash on August 1, 2026.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.