REVIEW 1 major objections 4 minor 13 references
Ping-pong in the projective plane over a nonarchimedean field
T0 review · 1 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash
Pith's one-line read Every lattice in SL(3) over a nonarchimedean field contains an undistorted Z^2 * Z subgroup.
desk verdict First discrete non-lattice subgroups in nonarchimedean SL_3, built by a clever ping-pong argument that has one unproved but likely fillable valuation bound. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing mechanism is a two-set ping-pong on the projective plane and its flag space. For $U = (1+\pi\mathfrak{m})^2$ inside the affine chart, and for each $x\in U$, let $V_x$ be the union of $U$ with every affine line through $x$ whose slope lies in $\pi+\pi\mathfrak{m}$. The paper's key combinatorial claim is that $V_x \cap \gamma U = \varnothing$ for every nontrivial $\gamma$ in the diagonal $\mathbb{Z}^2$ subgroup; this disjointness is what turns the action into a free product. A regular element $g$ whose attracting and repelling flags both lie in the associated open set $W$ then sends the complement of the two regions $V_{x_\pm}$ into $U$, so the ping-pong lemma applies. To control distortion, the paper uses an operator-norm pinning estimate (Proposition 6): the ping-pong inequalities are converted into a lower bound of the form $\|g\|\|g^{-1}\| \geq e^{\alpha |g| - C}$, which is the quantitative meaning of undistorted in this setting.
What would settle it
Compute, for a concrete nonarchimedean field such as $\mathbb{Q}_p$ or $\mathbb{F}_q((t))$, the slope $\sigma$ appearing in the last paragraph of the proof for a specific diagonal $\mathbb{Z}^2$ subgroup and specific $x,y\in U$; if any choice gives $\mathrm{val}_\pi(\sigma)=1$, then $V_x\cap \gamma U$ is nonempty and the free product decomposition claimed in Theorem 1 would fail for that configuration.
Extended reading notes
Core claim
At its heart the paper establishes Theorem 1: for every lattice $\Lambda < \mathrm{SL}_3(k)$ and every $\mathbb{Z}^2$ subgroup $\Delta' \subset \Lambda$, there is a finite-index subgroup $\Delta \subset \Delta'$ and an infinite-order element $g \in \Lambda$ such that $\langle \Delta, g\rangle$ is undistorted and decomposes as the free product $\Delta * \langle g\rangle$. The desired free factor $\Delta$ is a diagonal $\mathbb{Z}^2$ acting on the affine chart $\{Z\neq 0\}$ of $\mathbb{P}(k^3)$, with eigenvalues of norm less than $1$ in the relevant coordinates. The proof identifies an open set $W$ of projective flags $(x,L)$ in which $x$ lies near $(1,1)$ and the line $L$ has slope in $\pi + \pi\mathfrak{m}$; a regular element $h$ with attracting and repelling flags in $W$ then acts as the second ping-pong player after passing to a suitable power $g = h^{N_0}$. The paper also shows directly that the resulting subgroup is algebraically dense and infinite covolume, and argues that no group of the form $\mathbb{Z}^2 * \mathbb{Z}$ can be a lattice in any local field, so the construction really produces groups of a new discreteness type.
Load-bearing premise
The whole construction depends on the final, unproved assertion in the proof of Theorem 1 that the slope $\sigma$ of a certain line always satisfies $\sigma=\infty$ or $\mathrm{val}_\pi(\sigma)\neq 1$, because that valuation bound is what keeps the two ping-pong sets disjoint.
Editorial extensions
If this is right
- Corollary 3 follows: the lattice $\mathrm{SL}_3(\mathbb{F}_q[t])$ in $\mathrm{SL}_3(\mathbb{F}_q((1/t)))$ contains a subgroup isomorphic to $\mathbb{Z}^2 * \mathbb{Z}$.
- The subgroup is algebraically dense and of infinite covolume, so it provides the first examples of finitely generated algebraically dense infinite-covolume discrete subgroups of a nonarchimedean almost-simple group that are not virtually free.
- Because $\mathbb{Z}^2 * \mathbb{Z}$ cannot embed as a lattice in the $k$-points of any $k$-group, the theorem produces finitely generated discrete subgroups of nonarchimedean Lie groups that are not virtually isomorphic to any lattice.
- For $\mathrm{SL}_3(\mathbb{Z})$ the existence of a $\mathbb{Z}^2 * \mathbb{Z}$ subgroup remains open; the paper observes that a hypothetical such subgroup would act minimally on the flag boundary, so the ping-pong method used here cannot transfer to that case.
Reading between the lines
- The final step of the disjointness proof asserts, without proof, that the slope $\sigma$ of the line joining $\gamma x$ to $y$ always satisfies $\sigma=\infty$ or $\mathrm{val}_\pi(\sigma)\neq 1$; verifying this valuation bound, or finding a counterexample, is the natural next check. If the bound fails, the ping-pong sets could overlap and the free-product conclusion would not follow.
- The same slope-controlled construction could plausibly extend to other rank-two groups over nonarchimedean fields by replacing the condition 'slope in $\pi+\pi\mathfrak{m}$' with an analogous residue-condition, though the paper does not claim such an extension.
- The contrast with the open $\mathrm{SL}_3(\mathbb{Z})$ case suggests that the obstruction is boundary geometry rather than group theory: ultrametric valuation constraints can guarantee disjointness in a way that archimedean absolute values cannot.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proves (or aims to prove) that every lattice Λ in SL3(k), for a nonarchimedean local field k, contains an undistorted subgroup isomorphic to the free product Z^2 * Z. The proof uses a ping-pong argument on the Furstenberg boundary of SL3(k): it exhibits a finite-index subgroup Δ of a given Z^2 subgroup of Λ, a diagonal affine action on a chart, and a compact set U = (1+πm)^2 such that a certain set V_x of points and lines through x is disjoint from γU for every nontrivial γ∈Δ. A regular element g∈Λ with attracting and repelling flags in a related open set then gives the free product decomposition Δ∗⟨g⟩. The paper also contains a quantitative ping-pong lemma (Proposition 6) used to prove undistortedness, and remarks explaining Zariski density and why the constructed group is not virtually isomorphic to a lattice.
Significance. If the main theorem is correct, the paper provides the first known finitely generated discrete subgroups of nonarchimedean Lie groups that are not virtually isomorphic to lattices, in contrast to the open problem for SL3(Z). The overall strategy is natural and the paper is well structured: Proposition 6 is a clean, explicit ping-pong lemma with a convincing proof, and the reduction to the affine chart is geometrically reasonable. The construction is new and does not appear circular: the result is derived from standard external theorems about arithmetic lattices, tori, and Bruhat–Tits buildings. The central obstruction to accepting the proof as written is a single unproved valuation assertion, which is load-bearing for the ping-pong disjointness.
major comments (1)
- [Proof of Theorem 1, final sentence of the claim] The proof of the disjointness claim ends with the sentence 'The latter is true since either σ = ∞ or val_π(σ) ≠ 1', but no argument is supplied for this valuation assertion. This assertion is exactly what makes the ping-pong sets disjoint: if a line through x had slope in π+πm and met γU, the claimed free product decomposition would not follow. The text provides no case analysis for the signs of m and n, no computation of val_π(α) and val_π(β), and no treatment of how the factors 1+λ_i and 1+μ_i affect possible cancellations. A short case analysis may well prove the assertion, but as written it is a genuine gap in the proof of Theorem 1.
minor comments (4)
- [Abstract] The abstract as quoted at the beginning of the submission and the abstract printed in the body of the paper make different novelty claims: the former says 'not virtually isomorphic to lattices', while the latter says 'not virtually free'. These are not equivalent statements, and the paper should state its intended claim consistently, ideally matching the formulation supported by Remark 5.
- [Proof of Theorem 1, first paragraph of the claim] The sentence 'It is clear that U ∩ γU = ∅' is asserted without explanation. Since γ is an arbitrary nontrivial element of Δ and may have mixed expansion and contraction in the affine coordinates, a one-line valuation justification would improve the exposition.
- [Throughout] The displayed text contains many instances of the artifact '/integerdivide' where a set-difference or quotient symbol is clearly intended. These should be corrected in the final version so that formulas such as P(k^3) \ V_{x±} are readable.
- [Proof of Theorem 1, definition of σ] The reduction 'Note that β1^n, α2^m ∈ 1+πm' is stated without proof. It is plausible from the preceding normalization using p(q−1)-powers, but the authors should spell out why α2 and β1 become elements of 1+πm after that normalization.
Circularity Check
No circularity: the construction is derived from external theorems and a new ping-pong argument; the one questionable step is an unproved valuation bound, not a circular reduction.
full rationale
The paper's derivation chain does not reduce to its inputs. Theorem 1 is proved by constructing a finite-index subgroup Δ of a given Z^2 subgroup of a lattice, producing a ping-pong pair of sets, and invoking a folklore ping-pong lemma (Proposition 6) whose proof is given in the paper. The key disjointness claim V_x ∩ γU = ∅ is reduced to a slope computation, and the paper states: "The latter is true since either σ = ∞ or val_π(σ) ≠ 1." This sentence is a mathematical gap (the valuation bound is not proved), but it is not circular: it does not assume Theorem 1, it does not fit a parameter to the desired conclusion, and it does not redefine the conclusion. The slope σ is computed explicitly from the action of Δ, and the desired non-membership in π + πm is a consequence of the asserted valuation condition, not equivalent to it by construction. All external citations are to works by other authors: Margulis arithmeticity and normal subgroup theorem, Prasad–Rapinchuk for tori and regular elements, Bridson–Haefliger for Bruhat–Tits theory, Shalom for Zariski density, and others. There are no self-citations by Douba, Kubrak, or Tsouvalas that carry a load-bearing step. There is no fitted input renamed as a prediction, no uniqueness theorem imported from the authors' own prior work, and no known result repackaged under new coordinates. The construction of the undistorted free product is new and depends on explicit estimates in Proposition 6. The only concern raised by the proof is the unproved valuation assertion, which is a correctness risk rather than a circularity. Therefore the circularity score is 0.
Assumptions & free parameters
assumptions (5)
- domain assumption Margulis arithmeticity theorem: any lattice in SL3(k) is arithmetic.
- domain assumption Prasad-Rapinchuk Theorem 1(ii): existence of Z^2 subgroups and regular elements with prescribed flags.
- domain assumption Bruhat-Tits building theory: a discrete Z^2 subgroup preserves an apartment and acts cellularly.
- domain assumption Shalom's results on Zariski dense subgroups and invariant measures.
- domain assumption Bader-Furman-Sauer lattice envelopes, Margulis normal subgroup theorem, Prasad rigidity, Bass-Lubotzky tree lattices.
Cite this review
Pith. "Pith review of Ping-pong in the projective plane over a nonarchimedean field." pith.science (2026). https://pith.science/paper/5QFG6CGG
@misc{pith2026250513639,
author = {Pith},
title = {Pith review of: Ping-pong in the projective plane over a nonarchimedean field},
year = {2026},
howpublished = {\url{https://pith.science/paper/5QFG6CGG}},
note = {Machine review of arXiv:2505.13639}
}
abstract
We show that any lattice in $\mathrm{SL}_3(k)$, where $k$ is a nonarchimedean local field, contains an undistorted subgroup isomorphic to the free product $\mathbb{Z}^2*\mathbb{Z}$. To our knowledge, the subgroups we construct give the first examples in the literature of finitely generated discrete subgroups of nonarchimedean Lie groups that are not virtually isomorphic to lattices in such Lie groups. Our result is in contrast to the case of $\mathrm{SL}_3(\mathbb{Z})$, in which the existence of a $\mathbb{Z}^2*\mathbb{Z}$ subgroup remains open.
Reference graph
Works this paper leans on
- [1]
-
[2]
H. Bass and A. Lubotzky , Tree lattices, vol. 176 of Progress in Mathematics, Birkh¨ auser Boston, Inc., Boston, MA, 2001. With appendices by Bass, L. C arbone, Lubotzky, G. Rosen- berg and J. Tits
work page 2001
-
[3]
M. R. Bridson and A. Haefliger , Metric Spaces of Non-Positive Curvature , vol. 319, Springer, 1999
work page 1999
-
[4]
F. Bruhat and J. Tits , Groupes r´ eductifs sur un corps local, Inst. Hautes ´Etudes Sci. Publ. Math., (1972), pp. 5–251
work page 1972
-
[5]
Remarks on discrete subgroups with full limit sets in higher rank Lie groups
S. Dey and S. Hurtado , Remarks on discrete subgroups with full limit sets in higher rank Lie groups, arXiv preprint arXiv:2405.10209, (2024)
work page Pith review arXiv 2024
-
[6]
Helgason , Differential geometry, Lie groups, and symmetric spaces , vol
S. Helgason , Differential geometry, Lie groups, and symmetric spaces , vol. 80 of Pure and Applied Mathematics, Academic Press, Inc. [Harcourt Brace Jovanovich, Publishers], New York-London, 1978
work page 1978
-
[7]
M. Kapovich, A. Detinko, and A. Kontorovich , List of problems on discrete subgroups of Lie groups and their computational aspects , in Computational Aspects of Discrete Subgroups of Lie Groups, vol. 783 of Contemp. Math., Amer. Math. Soc., [ Providence], RI, [2023] ©2023, pp. 113–126
work page 2023
-
[8]
A. Lubotzky, S. Mozes, and M. S. Raghunathan , The word and Riemannian metrics on lattices of semisimple groups , Inst. Hautes ´Etudes Sci. Publ. Math., (2000), pp. 5–53 (2001)
work page 2000
Show all 13 references
-
[9]
G. A. Margulis , Discrete subgroups of semisimple Lie groups , vol. 17 of Ergebnisse der Mathematik und ihrer Grenzgebiete (3) [Results in Mathemat ics and Related Areas (3)], Springer-Verlag, Berlin, 1991
1991
-
[10]
Prasad , Discrete subgroups isomorphic to lattices in semisimple Li e groups , Amer
G. Prasad , Discrete subgroups isomorphic to lattices in semisimple Li e groups , Amer. J. Math., 98 (1976), pp. 241–261
1976
-
[11]
Prasad and A
G. Prasad and A. S. Rapinchuk , Irreducible tori in semisimple groups , Internat. Math. Res. Notices, (2001), pp. 1229–1242
2001
-
[12]
Shalom , Invariant measures for algebraic actions, Zariski dense su bgroups and Kazhdan ’s property (T), Trans
Y. Shalom , Invariant measures for algebraic actions, Zariski dense su bgroups and Kazhdan ’s property (T), Trans. Amer. Math. Soc., 351 (1999), pp. 3387–3412
1999
-
[13]
G. A. Soifer , Embedding Z2 ∗ Z as a discrete subgroup into small linear groups , Internat. J. Algebra Comput., 22 (2012), pp. 1250057, 10. Institut des Hautes ´Etudes Scientifiques, Universit ´e Paris-Saclay, 35 route de Chartres, 91440 Bures-sur-Yvette, France Email address ...
2012
Reviewed August 15, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.