Dense linear subhypergraph with quadratic edges

Conjecture 6.5 · arXiv:2401.00359

arXiv Conjecture high confidence— first stated 2023-12-31

Authors: Jacob Fox, Maya Sankar, Michael Simkin, Jonathan Tidor, Yunkun Zhou

Status open high confidence

Conjecture 6.5 from arXiv:2401.00359 asks for a constant c>0 such that every n-vertex 3-uniform hypergraph with at least n^{3-c/d} edges contains a linear subhypergraph with d vertices and at least d^2/100 edges. No resolution or substantial partial progress on this conjecture was found in the literature through May 2026. The authors note that the non-linear analogue is known (by a hypergraph Kővári–Sós–Turán argument), and the conjecture represents the key missing step toward proving ex(n,H_L) ≤ O(n^{3-c/d}) for linear 3-uniform hypergraphs H_L of bounded skeletal degeneracy.

Reviewer notes. No follow-up found after five web searches covering the source paper's arXiv page, the authors' recent output, and the relevant mathematical terms. The conjecture was posed in the concluding remarks of the paper as a key open problem; its resolution would require a new technique beyond the hypergraph Kővári–Sós–Turán argument (which handles only the non-linear version). The paper arXiv:2408.09029 ('An Improved Turán Exponent for 2-Complexes') by related authors appeared in August 2024 but could not be verified within the call cap as addressing this specific conjecture.

Auto-reviewed 2026-05-15 with claude-sonnet-4-6 (web search enabled).

Conjecture. There is a constant $c>0$ such that for all $d\geq 4$ the following holds. If $G$ is a (not necessarily linear) 3-uniform hypergraph on $n$ vertices with at least $n^{3-c/d}$ edges, then there is a linear subhypergraph of $G$ with $d$ vertices and at least $d^{2}/100$ edges.

Context

Proving that $\mathrm{ex}(n,H_{L})\leq O(n^{3-c/d})$ would require finding $H_{L}$ — a linear hypergraph — in any sufficiently dense host. The current barrier is that the techniques only find $K^{(3)}_{d,d,d}$, a non-linear structure. If Conjecture 6.5 holds, finding any dense enough linear subhypergraph (with quadratically many edges) would be the key missing step. The conjecture is known without the linearity constraint, since a hypergraph Kővári–Sós–Turán argument gives $K^{(3)}_{1,(d-1)/2,(d-1)/2}$ in any 3-uniform hypergraph with $n^{3-c/d}$ edges.

Source paper

Ramsey and Turán numbers of sparse hypergraphs
Jacob Fox, Maya Sankar, Michael Simkin, Jonathan Tidor, Yunkun Zhou · 2023-12-31
https://arxiv.org/abs/2401.00359