The relations of a Lie subalgebra considered by Tao: a complete presentation

Primarily AI-generated textHuman understanding: some partsmath.RA — Rings and Algebrasmath.CO — Combinatorics

Contributed by Qiyuan Ma ↗

Version 1 / Nov 17, 2026 / CC BY 4.0

Abstract

Let g be the Lie algebra over a field of characteristic 0 generated by D_i, E_i (i ≥ 1), subject to [D_i,D_j] = [E_i,E_j] = 0 and [D_i,E_j] = [D_j,E_i]. Let h be the subalgebra generated by the elements [D_i,E_j] and E_i. On MathOverflow (question 464808), T. Tao asked whether the obvious relations [D_i,E_j] = [D_j,E_i], [E_i,E_j] = 0 and [E_i,[D_j,E_k]] = [E_k,[D_j,E_i]] generate all relations of h. We prove that they do. The proof has four steps: (1) a polarisation argument reduces the problem to multilinear components; (2) we give a presentation of the ideal generated by the E_i; (3) an excess filtration, whose associated graded algebra we compute explicitly, reduces injectivity to an equality of dimensions; (4) that equality follows from the Hilbert series of g, which is given by Eulerian numbers. The Hilbert series in step 4 is equivalent to known results. The multilinear components of the enveloping algebra of the ideal generated by the E_i coincide with the homology of the Losev–Manin spaces: their Betti numbers are due to Losev and Manin (Michigan Math. J. 48 (2000)), and the algebra structure was presented by Dotsenko, Shadrin and Tamaroff (Comm. Math. Phys. 399 (2023)). The paper explains this relation precisely. It also gives an independent, self-contained derivation, using a PBW basis of U(g) obtained from Bergman’s diamond lemma and a sign-reversing involution. By-products: h is a semidirect product of a free Lie algebra and an abelian Lie algebra; the elements [D_i,E_j] (i ≤ j) freely generate a free Lie subalgebra of g; the multilinear components of g have dimensions given by Eulerian numbers (equivalent to known results). We were unable to find the affirmative answer to Tao's question or the freeness statement in the literature. Independent computer checks are included in the separate code archive.

Author comments: 17 pages, version 1.1. This work answers MathOverflow question 464808.

Provenance statement

The submitting human author, Qiyuan Ma, reports that some parts of this work have been checked and understood by the human author, but not all parts. The human author has not specified which parts were checked. AI assistance (Codex (OpenAI) and Claude (Anthropic); specific version numbers were not recorded) was used in the literature search, in designing and running the computational experiments, in writing the accompanying verification code, and in drafting the manuscript. AI is not an author. The human author is solely responsible for the work.

Tools used

OpenAI
Codex
Anthropic
Claude

References

  1. Jörgen Backelin and Ralf Fröberg, Koszul algebras, V eronese subrings and rings with linear resolutions , Rev. Roumaine Math. Pures Appl. 30 (1985), no. 2, 85--97.
  2. George M. Bergman, The diamond lemma for ring theory , Adv. Math. 29 (1978), no. 2, 178--218.
  3. Vladimir Dotsenko, Sergey Shadrin, and Pedro Tamaroff, Generalized cohomological field theories in the higher order formalism , Comm. Math. Phys. 399 (2023), no. 3, 1439--1500, doi:10.1007/s00220-022-04577-6 (https://doi.org/10.1007/s00220-022-04577-6), arXiv:2112.06015 (https://arxiv.org/abs/2112.06015).arXivDOI
  4. A. Losev and Yu. Manin, New moduli spaces of pointed curves and pencils of flat connections , Michigan Math. J. 48 (2000), no. 1, 443--472, doi:10.1307/mmj/1030132728 (https://doi.org/10.1307/mmj/1030132728), arXiv:math/0001003 (https://arxiv.org/abs/math/0001003).arXivDOI
  5. Terry Tao, Relations in a certain L ie algebra , MathOverflow, question 464808, February 24, 2024, https://mathoverflow.net/questions/464808.link
  6. YCor, comments on MO , February 24--27, 2024 (UTC), https://mathoverflow.net/posts/comments/1207321. The reformulation and the question about the elements [Di,Ej][D_i,E_j] are in this first comment.link

Version history

  1. v1 · 17 pagesSubmitted by Qiyuan Ma17 pages, version 1.1. This work answers MathOverflow question 464808.CurrentNov 17, 2026