Alec Kriebel

Two consequences of group-algebra and planar Fourier certificates: wreath products and triangle bounds

This note connects two substantial external results to questions in group theory and planar packing bounds. It gives explicit deductions and a supporting verification package. The two consequences have separate hypotheses and proofs.

The substantive inputs and prior work

The central inputs are OpenAI's A Torsion-Free Group Algebra That Is Not Directly Finite, Theorem 1.1, and A sharp Fourier certificate for planar circle packing, Theorem 1.1. These linked editions are fixed to the release used by the paper.

The wreath-product reduction is already part of Bradford–Fournier-Facio's framework; the planar product comparison is due to Cohn–de Laat–Salmon. Hopfian embedding and lattice Poisson arguments are established methods. The note presents attributed consequences, without claiming independent discovery or first priority.

Abstract

We derive two separately motivated consequences of precisely identified input theorems in OpenAI's mathematics collection. The torsion-free group-algebra counterexample, combined with the established Hopfian embedding framework, yields a finitely generated Hopfian group H for which the standard restricted regular wreath product C₂ wr H is non-Hopfian. We give its split epimorphism and kernel explicitly. The sharp planar two-point Fourier certificate, combined with the Cohn–de Laat–Salmon product-program comparison and triangular-lattice Poisson bounds, yields L₂ = 2/√3 and the exact planar triangle infimum P₂ = 4/3. We reconstruct the deductions, audit the substantive inputs, repair product-proof bookkeeping, and provide independently implemented validated computations and source records. These consequences use credited literature; no first-priority or independent-breakthrough claim is made. The planar equality does not settle the all-dimensional triangle problem or assert attainment of its infimum.

Hopfian factors with a non-Hopfian wreath product

A group is Hopfian when every surjective endomorphism is injective. The first consequence produces a finitely generated Hopfian group H for which the standard restricted regular wreath product C₂ wr H is non-Hopfian, even though both factors are finitely generated and Hopfian. Here C₂ is the two-element group. Its lamps are indexed by H, only finitely many are nonzero, and H acts by left translation.

The upstream group-algebra theorem supplies a finitely generated group with elements a,b satisfying ab = 1 but ba ≠ 1 over the two-element field. A classical embedding places this group inside a finitely generated Hopfian H, preserving that one-sided inverse defect. Put R = F₂[H], so the whole finite-support lamp base is the additive group of R. On Radd ⋊ H the explicit map and its section are

Φ(r,h) = (rb,h),   Σ(r,h) = (ra,h),   ΦΣ = id.

Right multiplication commutes with the left H-action. The map is onto, fixes H pointwise, and has the nontrivial kernel

ker Φ = R(1 − ba) × {1H}.

This answers the stated existence question in Kourovka Problem 21.22 and Bradford–Fournier-Facio Question 1.14 using the credited input. It concerns the restricted regular action and the C₂ lamp; it makes no assertion for a different action or a prescribed nonabelian lamp.

Exact planar two-point and triangle values

The second consequence concerns Fourier bounds for configurations with minimum distance one. With Fourier transform ĝ(ξ) = ∫g(x) exp(−2πi⟨x,ξ⟩) dx, let Ln be the infimum of g(0) over real Schwartz functions satisfying ĝ(0) = 1, ĝ ≥ 0, and g(x) ≤ 0 for |x| ≥ 1. Schwartz functions are smooth functions whose derivatives decay faster than every inverse power.

For the triangle program, use real Schwartz F on Rn × Rn with F̂(0,0) = 1 and F̂ ≥ 0. Require F(x,y) ≤ 0 whenever the three lengths |x|, |y|, |x − y| lie in D = {0} ∪ [1,∞), except at (0,0). Pn is the infimum of F(0,0). Zero coordinates, x = y and length-one boundaries are included; there is no additional |x + y| condition.

L₂ = 2/√3,   P₂ = 4/3,   √P₂ = L₂.

Triangular-lattice Poisson summation supplies matching lower bounds. OpenAI's planar two-point certificate supplies the sharp upper bound for L₂; the Cohn–de Laat–Salmon comparison Pn ≤ Ln2 then gives the planar triangle value. The paper sets out the comparison with consistent distance-one normalization.

These are raw number-density objectives, rather than covered-area density π/(2√3) for disks of radius 1/2. The conclusion is an equality of infima in dimension two: it does not settle the all-dimensional triangle problem, characterize triangle optimizers, assert their attainment, or give an objective-preserving conversion of every two-point optimizer.

Reading the evidence

The paper combines written arguments with exact and validated finite computations. Those computations support specific inequalities; they do not replace the group-theoretic, topological or infinite analytic arguments. Related Lean artifacts concern another construction and do not formally verify these consequences.

Paper and supporting materials

Open the unchanged published PDF · Download PDF · Zenodo original PDF

Download the original source and verification ZIP · Zenodo original ZIP · Download BibTeX

The ZIP contains the editable source, bibliography, dependency and theorem ledgers, exact source references, supporting code, evidence and reproduction instructions. The Zenodo record preserves the archived release.

Selected scholarly sources

The manuscript also discusses Daria Zhitniaia's institutional thesis abstract and J. Zhou's earlier preprint as limits on priority claims. It does not claim a complete comparison with the unavailable thesis or establish firstness. The full bibliography is in the PDF and ZIP.

Citation and rights

Alec Kriebel. Two consequences of group-algebra and planar Fourier certificates: wreath products and triangle bounds. Version 1.0.0, 6 October 2026. Zenodo. doi:10.5281/zenodo.23203143.

The preprint, authored proofs, exposition and JSON data are available under CC BY 4.0. Newly authored Python code in the supporting package is under its included MIT licence. Third-party work retains its original authorship and rights.

Contributions and collaboration

Independent verification of the inputs and deductions, or substantive work toward a journal version, would be valuable. Authorship would reflect actual mathematical contributions.

AI use and publication status

AI tools were used extensively in research, drafting, computation and verification; internal reviews were AI-assisted. This is an unrefereed preprint and has not undergone conventional human peer review. AI systems are not authors, and no external endorsement is implied.