Mendel–Naor sharp metric cotype question for L1

For 1≤p<∞1\le p<\infty, even m≥4m\ge 4, and n∈Nn\in\mathbb N, let Ap(n,m)A_p(n,m) be the least constant such that every function f:Zmn→L1f:\mathbb Z_m^n\to L_1 satisfies

∑i=1nEx∈Zmn∥f(x+m2ei)−f(x)∥1p≤Ap(n,m) Ex∈Zmn, ε∈{−1,1}n∥f(x+ε)−f(x)∥1p.\sum_{i=1}^n\mathbb E_{x\in\mathbb Z_m^n}\left\|f\left(x+\frac m2 e_i\right)-f(x)\right\|_1^p\le A_p(n,m)\,\mathbb E_{x\in\mathbb Z_m^n,\,\varepsilon\in\{-1,1\}^n}\left\|f(x+\varepsilon)-f(x)\right\|_1^p.

Determine the optimal order of Ap(n,m)A_p(n,m). The conjectured sharp answer is

Ap(n,m)1/p≍(mpn(1−p/2)++n)1/p,A_p(n,m)^{1/p}\asymp\left(m^p n^{(1-p/2)_+}+n\right)^{1/p},

with comparison constants independent of pp, mm, and nn. In particular, for p=2p=2 the desired order is A2(n,m)≍m2+nA_2(n,m)\asymp m^2+n, yielding the sharp metric-cotype scale m=O(n)m=O(\sqrt n) for L1L_1.

References

Progress summary

Refreshed
Claimed solved

A September 2026 unrefereed preprint claims to settle the sharp metric-cotype scale for L1L_1, but the claim has not been independently verified.

The problem asks for the optimal order in the torus inequality governing metric cotype for L1L_1, including whether the conjectural quadratic scale can be attained.

Known results

  • Earlier work recorded L1L_1 as unresolved, with the best bound m≳n3/2m\gtrsim n^{3/2}; the conjectured optimum was m≥n1/qm\ge n^{1/q}, hence m≥nm\ge\sqrt n for cotype q=2q=2.

September 8, 2026 claimed sharp bound

The preprint Sharp Metric Cotype Inequalities for L1L_1 via Nonlinear Cut Smoothing claims the optimal torus-inequality order, including the quadratic case with scaling m=O(n)m=O(\sqrt n). This would settle the tracked question, but the result is explicitly unrefereed and remains unverified.

Current status (as of September 2026): The conjectural sharp scale m=O(n)m=O(\sqrt n) is claimed in an unrefereed preprint, but independent verification is absent.

Sources

Solutions 0

No solutions have been posted yet.