An ErdH{o}s Matching Conjecture for Vector Spaces
Pith reviewed 2026-06-25 23:16 UTC · model grok-4.3
The pith
A vector-space analogue of the Erdős matching conjecture holds for k=2, for n exactly (s+1)k, and for all sufficiently large n.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
The conjecture asserts that m_q(n,k,s) equals the maximum of the Gaussian binomial [(s+1)k-1 choose k]_q and [n choose k]_q minus q^{ks} times [n-s choose k]_q for every n at least (s+1)k. This equality is established when k=2 (a vector-space Erdős-Gallai theorem), when n=(s+1)k, and when n is sufficiently large, via Lovász's minimax theorem for matroid matchings together with a high-dimensional Hoffman bound.
What carries the argument
The two constructions (all k-subspaces inside a fixed ((s+1)k-1)-dimensional subspace, and all k-subspaces meeting a fixed s-dimensional subspace nontrivially) together with Lovász's minimax theorem for matroid matchings and the high-dimensional Hoffman bound on uniform hypergraphs.
If this is right
- The conjecture holds for every choice of parameters when k equals 2.
- The conjecture holds for every choice of parameters when n equals (s+1)k.
- The conjecture holds for every choice of parameters when n is sufficiently large.
- A Hilton-Milner-type stability theorem identifies the largest families that are not one of the two constructions when n is large.
- The maximum size of a t-cover-free family of subspaces is determined up to a lower-order term for every t.
Where Pith is reading between the lines
- The same two constructions may remain extremal for the remaining open parameter ranges.
- The matroid-matching and Hoffman-bound techniques may resolve the conjecture for additional fixed values of k.
Load-bearing premise
The two listed constructions are the only ones that achieve the maximum size.
What would settle it
A collection of k-subspaces larger than both constructions for some parameters with k greater than 2 and n only moderately larger than (s+1)k.
read the original abstract
We study a vector-space analogue of the Erd\H{o}s Matching Conjecture. Let $m_q(n,k,s)$ denote the maximum cardinality of a family of $k$-dimensional subspaces of an $n$-dimensional vector space over $\mathbb F_q$ with no $s+1$ members whose sum is direct. Two natural constructions provide lower bounds. The first consists of all $k$-subspaces contained in a fixed $((s+1)k-1)$-dimensional subspace; the second consists of all $k$-subspaces that intersect a fixed $s$-dimensional subspace nontrivially. These constructions motivate the following vector-space analogue of the Erd\H{o}s Matching Conjecture: for all $n\ge (s+1)k$, $$m_q(n,k,s)=\max\left\{\genfrac{[}{]}{0pt}{}{(s+1)k-1}{k}_q,~\genfrac{[}{]}{0pt}{}{n}{k}_q-q^{ks}\genfrac{[}{]}{0pt}{}{n-s}{k}_q\right\}.$$ We prove this conjecture when $k=2$, when $n=(s+1)k$, and when $n$ is sufficiently large. In particular, the case $k=2$ may be viewed as a vector-space analogue of the Erd\H{o}s--Gallai theorem. In the large-$n$ range, we also prove a Hilton--Milner-type stability theorem, determining the largest nontrivial families with this property. Finally, we connect this problem with $t$-cover-free families in vector spaces and determine their extremal number up to a lower-order term, extending a recent result of Shan and Zhou for the special case $t=2$. The proofs combine Lov\'asz's minimax theorem for matroid matchings, a high-dimensional Hoffman bound for uniform hypergraphs, and packing-design arguments in vector spaces.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper defines m_q(n,k,s) as the maximum size of a family of k-dimensional subspaces of an n-dimensional vector space over F_q with no s+1 members whose sum is direct. It proposes the conjecture that for n ≥ (s+1)k this equals the maximum of two quantities arising from the 'all k-subspaces in a fixed ((s+1)k-1)-space' and 'all k-subspaces meeting a fixed s-space nontrivially' constructions. The manuscript proves the equality when k=2, when n=(s+1)k, and when n is sufficiently large; it also proves a Hilton-Milner-type stability theorem for large n and determines the extremal function for t-cover-free families in vector spaces up to lower-order terms, extending Shan-Zhou. The proofs combine Lovász's minimax theorem for matroid matchings, a high-dimensional Hoffman bound on an intersection hypergraph, and vector-space packing arguments.
Significance. If the derivations hold, the work is a substantive contribution to q-analogues of extremal set theory. It establishes the vector-space Erdős Matching Conjecture in three nontrivial regimes (including the k=2 case as a direct analogue of the Erdős-Gallai theorem) and supplies a stability result together with an application to t-cover-free families. The combination of matroid matching, spectral bounds, and packing arguments is appropriate and yields explicit extremal constructions that match the claimed upper bounds in each regime treated.
minor comments (3)
- [Introduction] The introduction should include an explicit definition of the q-binomial coefficient [n choose k]_q (including the product formula) before its first use in the conjecture statement, for accessibility to readers outside the q-analogue literature.
- [Section on large-n case] In the large-n regime, the manuscript should state the explicit lower bound on n (in terms of s, k, q) that makes the packing argument and Hoffman bound yield the claimed equality, rather than leaving 'sufficiently large' unqualified.
- [Stability theorem] The stability theorem for large n is stated only for the maximum nontrivial families; a brief remark on whether the same packing analysis yields stability for the k=2 or n=(s+1)k cases would strengthen the presentation.
Simulated Author's Rebuttal
We thank the referee for their thorough reading and positive evaluation of the manuscript. The recommendation for minor revision is noted, and we are pleased that the combination of methods and the results in the k=2, n=(s+1)k, and large-n regimes are viewed as substantive contributions. Since the report lists no specific major comments requiring clarification or correction, we have no point-by-point revisions to address at this stage.
Circularity Check
No significant circularity
full rationale
The paper states a vector-space analogue of the Erdős Matching Conjecture motivated by two explicit constructions, then proves the equality holds for k=2, n=(s+1)k, and sufficiently large n. The proofs rely on Lovász's minimax theorem for matroid matchings, the high-dimensional Hoffman bound on intersection hypergraphs, and vector-space packing arguments; these are external, independently established results applied directly to the linear matroid over F_q. No step reduces the claimed bound to a fitted parameter, a self-referential definition, or a load-bearing self-citation chain. The stability result and the connection to t-cover-free families likewise invoke external prior work without circular reduction.
Axiom & Free-Parameter Ledger
axioms (2)
- standard math Lovász's minimax theorem for matroid matchings
- standard math High-dimensional Hoffman bound for uniform hypergraphs
Reference graph
Works this paper leans on
-
[1]
On the size of graphs with complete-factors.Journal of Graph Theory, 9(1):197–201, 1985
Jin Akiyama and Peter Frankl. On the size of graphs with complete-factors.Journal of Graph Theory, 9(1):197–201, 1985
1985
-
[2]
Sur le couplage maximum d’un graphe.Comptes Rendus Hebdomadaires des S´ eances de l’Acad´ emie des Sciences, 247:258–259, 1958
Claude Berge. Sur le couplage maximum d’un graphe.Comptes Rendus Hebdomadaires des S´ eances de l’Acad´ emie des Sciences, 247:258–259, 1958
1958
-
[3]
Blackburn and Tuvi Etzion
Simon R. Blackburn and Tuvi Etzion. The asymptotic behavior of Grassmannian codes.IEEE Transactions on Information Theory, 58(10):6605–6609, 2012
2012
-
[4]
Brouwer, Ameera Chowdhury, Peter Frankl, T
Aart Blokhuis, Andries E. Brouwer, Ameera Chowdhury, Peter Frankl, T. Mussche, Bal´ azs Patk´ os, and Tam´ as Sz¨ onyi. A Hilton–Milner theorem for vector spaces.Electronic Journal of Combina- torics, 17(1):R71, 2010
2010
-
[5]
Bollob´ as, David E
B. Bollob´ as, David E. Daykin, and Paul Erd˝ os. Sets of independent edges of a hypergraph.The Quarterly Journal of Mathematics, 27(1):25–32, 1976
1976
-
[6]
Brouwer, Arjeh M
Andries E. Brouwer, Arjeh M. Cohen, and Arnold Neumaier.Distance-Regular Graphs, volume 18 ofErgebnisse der Mathematik und ihrer Grenzgebiete. 3. Folge. Springer, Berlin, 1989
1989
-
[7]
Mengyu Cao, Mei Lu, Benjian Lv, and Kaishun Wang.r-crosst-intersecting families for vector spaces.Journal of Combinatorial Theory, Series A, 193:105688, 2023
2023
-
[8]
Shadows and intersections in vector spaces.Journal of Combinatorial Theory, Series A, 117(8):1095–1106, 2010
Ameera Chowdhury and Bal´ azs Patk´ os. Shadows and intersections in vector spaces.Journal of Combinatorial Theory, Series A, 117(8):1095–1106, 2010
2010
-
[9]
PhD thesis, University of California, San Diego, 2012
Ameerah Naz Chowdhury.Shadows and Intersections. PhD thesis, University of California, San Diego, 2012. 22
2012
-
[10]
Association schemes andt-designs in regular semilattices.Journal of Combi- natorial Theory, Series A, 20(2):230–243, 1976
Philippe Delsarte. Association schemes andt-designs in regular semilattices.Journal of Combi- natorial Theory, Series A, 20(2):230–243, 1976
1976
-
[11]
A problem on independentr-tuples.Annales Universitatis Scientiarum Budapesti- nensis de Rolando E¨ otv¨ os Nominatae
Paul Erd˝ os. A problem on independentr-tuples.Annales Universitatis Scientiarum Budapesti- nensis de Rolando E¨ otv¨ os Nominatae. Sectio Mathematica, 8:93–95, 1965
1965
-
[12]
Families of finite sets in which no set is covered by the union of two others.Journal of Combinatorial Theory, Series A, 33(2):158–166, 1982
Paul Erd˝ os, Peter Frankl, and Zolt´ an F¨ uredi. Families of finite sets in which no set is covered by the union of two others.Journal of Combinatorial Theory, Series A, 33(2):158–166, 1982
1982
-
[13]
Families of finite sets in which no set is covered by the union ofrothers.Israel Journal of Mathematics, 51(1–2):79–89, 1985
Paul Erd˝ os, Peter Frankl, and Zolt´ an F¨ uredi. Families of finite sets in which no set is covered by the union ofrothers.Israel Journal of Mathematics, 51(1–2):79–89, 1985
1985
-
[14]
On maximal paths and circuits of graphs.Acta Mathematica Academiae Scientiarum Hungaricae, 10(3–4):337–356, 1959
Paul Erd˝ os and Tibor Gallai. On maximal paths and circuits of graphs.Acta Mathematica Academiae Scientiarum Hungaricae, 10(3–4):337–356, 1959
1959
-
[15]
Paul Erd˝ os, Chao Ko, and R. Rado. Intersection theorems for systems of finite sets.The Quarterly Journal of Mathematics, 12(1):313–320, 1961
1961
-
[16]
Problems onq-analogs in coding theory.arXiv preprint arXiv:1305.6126, 2013
Tuvi Etzion. Problems onq-analogs in coding theory.arXiv preprint arXiv:1305.6126, 2013
Pith/arXiv arXiv 2013
-
[17]
Grassmannian codes with new distance measures for network coding
Tuvi Etzion and Hui Zhang. Grassmannian codes with new distance measures for network coding. IEEE Transactions on Information Theory, 65(7):4131–4142, 2019
2019
-
[18]
High-dimensional Hoffman bound and applications in extremal combinatorics.Algebraic Combinatorics, 4(6):1005–1026, 2021
Yuval Filmus, Konstantin Golubev, and Noam Lifshitz. High-dimensional Hoffman bound and applications in extremal combinatorics.Algebraic Combinatorics, 4(6):1005–1026, 2021
2021
-
[19]
The shifting technique in extremal set theory
Peter Frankl. The shifting technique in extremal set theory. In C. Whitehead, editor,Surveys in Combinatorics 1987, volume 123 ofLondon Mathematical Society Lecture Note Series, pages 81–110. Cambridge University Press, 1987
1987
-
[20]
Improved bounds for Erd˝ os’ matching conjecture.Journal of Combinatorial Theory, Series A, 120(5):1068–1072, 2013
Peter Frankl. Improved bounds for Erd˝ os’ matching conjecture.Journal of Combinatorial Theory, Series A, 120(5):1068–1072, 2013
2013
-
[21]
On the maximum number of edges in a hypergraph with given matching number
Peter Frankl. On the maximum number of edges in a hypergraph with given matching number. Discrete Applied Mathematics, 216:562–581, 2017
2017
-
[22]
Proof of the Erd˝ os matching conjecture in a new range.Israel Journal of Mathe- matics, 222(1):421–430, 2017
Peter Frankl. Proof of the Erd˝ os matching conjecture in a new range.Israel Journal of Mathe- matics, 222(1):421–430, 2017
2017
-
[23]
On non-trivial families without a perfect matching.European Journal of Combina- torics, 84:103044, 2020
Peter Frankl. On non-trivial families without a perfect matching.European Journal of Combina- torics, 84:103044, 2020
2020
-
[24]
Colored packing of sets
Peter Frankl and Zolt´ an F¨ uredi. Colored packing of sets. In C. J. Colbourn and R. Mathon, editors,Combinatorial Design Theory, volume 149 ofNorth-Holland Mathematics Studies, pages 165–177. North-Holland, Amsterdam, 1987
1987
-
[25]
Peter Frankl and Ronald L. Graham. Intersection theorems for vector spaces.European Journal of Combinatorics, 6(2):183–187, 1985
1985
-
[26]
The Erd˝ os matching conjecture and concentration inequalities
Peter Frankl and Andrey Kupavskii. The Erd˝ os matching conjecture and concentration inequalities. Journal of Combinatorial Theory, Series B, 157:366–400, 2022
2022
-
[27]
Peter Frankl, Hongliang Lu, Jie Ma, and Yuze Wu. Towards the Erd˝ os matching conjecture for 4-uniform hypergraphs: Stability and applications.arXiv preprint arXiv:2602.19230, 2026. 23
Pith/arXiv arXiv 2026
-
[28]
On matchings in hypergraphs.Elec- tronic Journal of Combinatorics, 19(2):P42, 2012
Peter Frankl, Tomasz Luczak, and Katarzyna Mieczkowska. On matchings in hypergraphs.Elec- tronic Journal of Combinatorics, 19(2):P42, 2012
2012
-
[29]
Near perfect coverings in graphs and hypergraphs.European Journal of Combinatorics, 6(4):317–326, 1985
Peter Frankl and Vojtech R¨ odl. Near perfect coverings in graphs and hypergraphs.European Journal of Combinatorics, 6(4):317–326, 1985
1985
-
[30]
On the maximum number of edges in a triple system not containing a disjoint family of a given size.Combinatorics, Probability and Computing, 21(1–2):141–148, 2012
Peter Frankl, Vojtech R¨ odl, and Andrzej Ruci´ nski. On the maximum number of edges in a triple system not containing a disjoint family of a given size.Combinatorics, Probability and Computing, 21(1–2):141–148, 2012
2012
-
[31]
Peter Frankl and Richard M. Wilson. The Erd˝ os–Ko–Rado theorem for vector spaces.Journal of Combinatorial Theory, Series A, 43(2):228–236, 1986
1986
-
[32]
Christopher Godsil and Karen Meagher.Erd˝ os–Ko–Rado Theorems: Algebraic Approaches, volume
-
[33]
Cambridge University Press, 2015
2015
-
[34]
Kleitman
Curtis Greene and Daniel J. Kleitman. Proof techniques in the theory of finite sets. In Gian-Carlo Rota, editor,Studies in Combinatorics, volume 17 ofMAA Studies in Mathematics, pages 22–79. Mathematical Association of America, Washington, D.C., 1978
1978
-
[35]
A stability result for almost perfect matchings.Journal of Graph Theory, 112(1):5–14, 2026
Mingyang Guo and Hongliang Lu. A stability result for almost perfect matchings.Journal of Graph Theory, 112(1):5–14, 2026
2026
-
[36]
Anthony J. W. Hilton and Eric C. Milner. Some intersection theorems for systems of finite sets. The Quarterly Journal of Mathematics, 18(1):369–384, 1967
1967
-
[37]
Alan J. Hoffman. On eigenvalues and colorings of graphs. In Charles A. Micchelli, editor,Selected Papers of Alan J. Hoffman: With Commentary, pages 407–419. World Scientific, 2003
2003
-
[38]
Jianfeng Hou, Caiyun Hu, and Xizhi Liu. A finite-board reduction for the Erd˝ os matching conjec- ture and the 4-uniform case via exact certificates.arXiv preprint arXiv:2605.26060, 2026
Pith/arXiv arXiv 2026
-
[39]
Intersection theorems for systems of finite vector spaces.Discrete Mathematics, 12(1):1–16, 1975
Wen-Ning Hsieh. Intersection theorems for systems of finite vector spaces.Discrete Mathematics, 12(1):1–16, 1975
1975
-
[40]
The size of a hypergraph and its matching number
Hao Huang, Po-Shen Loh, and Benny Sudakov. The size of a hypergraph and its matching number. Combinatorics, Probability and Computing, 21(3):442–450, 2012
2012
-
[41]
Degree versions of the Erd˝ os–Ko–Rado theorem and Erd˝ os hypergraph matching conjecture.Journal of Combinatorial Theory, Series A, 150:233–247, 2017
Hao Huang and Yi Zhao. Degree versions of the Erd˝ os–Ko–Rado theorem and Erd˝ os hypergraph matching conjecture.Journal of Combinatorial Theory, Series A, 150:233–247, 2017
2017
-
[42]
A survey of cover-free families: Constructions, applica- tions, and generalizations
Tha´ ıs Bardini Idalino and Lucia Moura. A survey of cover-free families: Constructions, applica- tions, and generalizations. InProceedings of the International Conference on New Advances in Designs, Codes and Cryptography, pages 195–239. Springer, 2022
2022
-
[43]
Remarks on the Erd˝ os matching conjecture for vector spaces.European Journal of Combinatorics, 94:103306, 2021
Ferdinand Ihringer. Remarks on the Erd˝ os matching conjecture for vector spaces.European Journal of Combinatorics, 94:103306, 2021
2021
-
[44]
Structure oft-intersecting families of vector spaces
Ferdinand Ihringer and Andrey Kupavskii. Structure oft-intersecting families of vector spaces. arXiv preprint arXiv:2605.02698, 2026
Pith/arXiv arXiv 2026
-
[45]
G. O. H. Katona. A simple proof of the Erd˝ os–Chao Ko–Rado theorem.Journal of Combinatorial Theory, Series B, 13(2):183–184, 1972. 24
1972
-
[46]
W. H. Kautz and R. C. Singleton. Nonrandom binary superimposed codes.IEEE Transactions on Information Theory, 10(4):363–377, 1964
1964
-
[47]
Kleitman
Daniel J. Kleitman. Maximal number of subsets of a finite set nokof which are pairwise disjoint. Journal of Combinatorial Theory, 5(2):157–163, 1968
1968
-
[48]
Erd˝ os matching conjecture for almost perfect matchings
Dmitriy Kolupaev and Andrey Kupavskii. Erd˝ os matching conjecture for almost perfect matchings. Discrete Mathematics, 346(4):113304, 2023
2023
-
[49]
Isodiametric inequality for vector spaces.arXiv preprint arXiv:2503.19239, 2025
Jiaqi Liao, Hong Liu, and Guiying Yan. Isodiametric inequality for vector spaces.arXiv preprint arXiv:2503.19239, 2025
arXiv 2025
-
[50]
Jiuqiang Liu, Guihai Yu, Lihua Feng, and Yongtao Li.L-intersecting or configuration-forbidden families on set systems and vector spaces over finite fields.arXiv preprint arXiv:2403.04289, 2024
arXiv 2024
-
[51]
Matroid matching and some applications.Journal of Combinatorial Theory, Series B, 28(2):208–236, 1980
L´ aszl´ o Lov´ asz. Matroid matching and some applications.Journal of Combinatorial Theory, Series B, 28(2):208–236, 1980
1980
-
[52]
Selecting independent lines from a family of lines in a space.Acta Scientiarum Mathematicarum, 42(1–2):121–131, 1980
L´ aszl´ o Lov´ asz. Selecting independent lines from a family of lines in a space.Acta Scientiarum Mathematicarum, 42(1–2):121–131, 1980
1980
-
[53]
On Erd˝ os’ extremal problem on matchings in hy- pergraphs.Journal of Combinatorial Theory, Series A, 124:178–194, 2014
Tomasz Luczak and Katarzyna Mieczkowska. On Erd˝ os’ extremal problem on matchings in hy- pergraphs.Journal of Combinatorial Theory, Series A, 124:178–194, 2014
2014
-
[54]
The formula for Tur´ an number of spanning linear forests.Discrete Mathematics, 343(8):111924, 2020
Bo Ning and Jian Wang. The formula for Tur´ an number of spanning linear forests.Discrete Mathematics, 343(8):111924, 2020
2020
-
[55]
Covering Grassmannian codes: Bounds and constructions.IEEE Transactions on Information Theory, 69(6):3748–3758, 2023
Bingchen Qian, Xin Wang, Chengfei Xie, and Gennian Ge. Covering Grassmannian codes: Bounds and constructions.IEEE Transactions on Information Theory, 69(6):3748–3758, 2023
2023
-
[56]
On cover-free families of finite vector spaces.Journal of Combi- natorial Theory, Series A, 220:106131, 2026
Yunjing Shan and Junling Zhou. On cover-free families of finite vector spaces.Journal of Combi- natorial Theory, Series A, 220:106131, 2026. A Parameter calculations for Theorem 1.3(iii) We write [a]q := a 1 q. The following elementary estimate will be used repeatedly. Lemma A.1.Letm, r, tbe positive integers withm≥t+r. Then 1−4q −(m−t−r+2) [t]q m−1 r−1 q...
2026
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.