REVIEW 2 major objections 6 minor 14 references
A uniform Tits alternative for endomorphisms of the projective line
T0 review · 2 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash
Pith's one-line read Two polarized endomorphisms with distinct preperiodic sets generate a free semigroup, and exponential-growth subsemigroups of End(P1) have uniform exponential growth.
desk verdict A genuinely new uniform Tits alternative for End(P1), with a clean height-ping-pong core; the only real soft spot is an under-specified citation in Lemma 3.1 that a referee must verify. 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 object is the complete metric space $\mathcal{H}_L$ of functions on $V(\bar{K})$ at bounded distance from a fixed Weil height $h_L$, together with the contraction $\alpha_f(h) = d^{-1} f^* h$ for each endomorphism $f$ polarized by $L$ of degree $d$. The unique fixed point of $\alpha_f$ is the canonical height $h_f$, whose zero set is exactly $\mathrm{PrePer}(f)$. The other central tool is Proposition 2.1, a ping-pong lemma for injective contractions on a complete metric space: if two contractions have distinct fixed points and contraction ratios $c_1 + c_2 \leq 1$, then the semigroup they generate is free. The proof uses the attractor of the two-map iterated function system: if the attractor is disconnected, ordinary ping-pong on coding cylinders gives freeness; if it is connected, a Hausdorff-measure argument shows the two pieces intersect in measure zero, giving a measurable ping-pong. The definitions of diameter of independence $\Delta(F)$ and algebraic entropy $\Sigma(F)$ convert this freeness into uniform growth bounds through the inequality $\Sigma(F) \geq \log(2)/\Delta(F)$.
What would settle it
Search for a nontrivial word relation between $f(z)=z^2$ and $g(z)=z^2-2$ over $\mathbb{Q}$; Theorem 1.1 predicts that these two degree-two maps generate a free semigroup, so any identity of finite words in $f$ and $g$ would refute the main claim.
Extended reading notes
Core claim
The discovery is that equality or difference of preperiodic-point sets controls freeness. Theorem 1.1 states that if $f_1$ and $f_2$ are endomorphisms of a projective variety, both polarized by the same ample line bundle, and $\mathrm{PrePer}(f_1) \neq \mathrm{PrePer}(f_2)$, then $f_1$ and $f_2$ generate a free semigroup of rank 2. The proof attaches to each $f_i$ a canonical height $h_{f_i}$, the unique fixed point of the contraction $\alpha_i(h) = d_i^{-1} f_i^* h$ on the space of height functions, and shows $h_{f_1} \neq h_{f_2}$ precisely because their zero loci differ. Since each $d_i \geq 2$, the two contraction ratios sum to at most 1, so a ping-pong lemma on attractors applies and the semigroup is free. The uniform consequence for $\mathrm{End}(\mathbb{P}^1)$ in characteristic zero is the dichotomy: a finitely generated subsemigroup is either of polynomial growth or has finite diameter of independence, and therefore any exponential-growth semigroup has uniform exponential growth.
Load-bearing premise
The argument stands on an imported rigidity theorem from arithmetic dynamics that the paper does not prove: if the preperiodic points of one map form a Zariski-dense subset of the preperiodic points of another, then the two sets are actually equal. The proof cites this result rather than stating its precise hypotheses, so all of the main theorems inherit whatever conditions that result requires.
Editorial extensions
If this is right
- If two polarized endomorphisms sharing a line bundle have distinct preperiodic sets, they are independent, so any finite generating set containing them has diameter of independence at most 2.
- In any finitely generated subsemigroup of $\mathrm{End}(\mathbb{P}^1)$ over characteristic 0, non-polynomial growth implies a finite diameter of independence, so there is no intermediate growth: every such semigroup is either polynomially growing or exponentially growing.
- For semigroups containing a degree-at-least-two element, the algebraic entropy is bounded below by $\log(2)/2$, independently of the particular maps.
- The degree-one case reduces to linear groups in $\mathrm{PGL}_2$, where exponential growth forces a finite, though not uniform, bound on the diameter of independence.
Reading between the lines
- Because Theorem 1.1 is stated for projective varieties of any dimension, the height-ping-pong mechanism itself is not special to the projective line; the one-dimensional input only appears when Theorem 1.3 invokes the previous result that a degree-at-least-two subsemigroup contains two elements with distinct preperiodic sets.
- A concrete test would be to search for word relations between $f(z)=z^2$ and $g(z)=z^2-2$ over $\mathbb{Q}$; the theorem predicts the semigroup is free, so any finite relation found would immediately refute the main claim.
- The sharpness example in Remark 2.2, with contractions of ratio $c_n$ approaching $1/2$ from above that do satisfy a relation, suggests that the ratio bound $c_1+c_2 \leq 1$ is exactly what makes the freeness conclusion robust, and that the degree bound $d \geq 2$ for polarized endomorphisms is the arithmetic-dynamics analogue of this threshold.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proves a uniform version of the Tits alternative for finitely generated subsemigroups of End(P1). The main theorems are: Theorem 1.1, a height-function ping-pong result showing that two endomorphisms polarized by the same line bundle with different preperiodic sets generate a free semigroup of rank two; Theorem 1.2, a uniform diameter bound Δ(S) ≤ 2 for semigroups of endomorphisms semi-polarized by a common line bundle and containing two degree-at-least-two maps with different preperiodic sets; and Theorem 1.3, the special case over P1 in characteristic zero, which yields the dichotomy between polynomial growth and finite diameter of independence, hence uniform exponential growth for semigroups of exponential growth. The proof strategy is a contraction ping-pong on the space of height functions, together with a comparison lemma for preperiodic sets of semigroups of endomorphisms.
Significance. If the results hold, they answer Question 5.1 of [BHPT24] and strengthen the earlier non-uniform Tits alternative to a uniform one, with a dimension-free bound Δ(S) ≤ 2 in the non-linear case. The contraction ping-pong on canonical heights is clean, parameter-free, and gives a genuinely new proof of the uniform bound. The paper is concise and mostly well written. Its main external dependence is on arithmetic-dynamical rigidity results of Yuan–Zhang and Carney; these are cited but not stated, and the precise way they are used is the main point that needs to be made explicit.
major comments (2)
- [Section 3, Lemma 3.1] The step “apply [YZ21, Theorem 1.3, (3)⇒(1)], or [Car20, Theorem A, (3)⇒(4)] … to conclude that P = PrePer(w)” is the central rigidity input of the paper, but neither the statement of those theorems nor the verification of their hypotheses is supplied. The argument needs the implication: if P is a Zariski-dense subset of V and P ⊂ PrePer(f) ∩ PrePer(w), then PrePer(f) = PrePer(w). Please state the exact form of the cited results and explain why the hypotheses (field, polarization, degree, common iterate, characteristic, or any other condition) are satisfied at that point of Lemma 3.1. Because Lemma 3.1 feeds directly into Theorem 1.2 and Theorem 1.3, this is load-bearing; if the cited theorems only give equality under stronger hypotheses, the proof of the main results is incomplete.
- [Section 3, Lemma 3.1] The sentence “so by the Northcott property, all w-orbits in P are finite” is not justified by the Northcott property alone as written, since a naive degree bound [K(w^n(x)):K] ≤ d_w^n [K(x):K] grows with n. The step is nevertheless correct, because w is defined over the finitely generated field K, so w^n(x) ∈ V(K(x)) and hence [K(w^n(x)):K] ≤ [K(x):K] is bounded. Please add this one-line explanation; with it, the conclusion P ⊂ PrePer(w) follows. As it stands, the printed proof gives the reader an unnecessary obstacle at a central point.
minor comments (6)
- [Proof of Theorem 1.1] The sentence about preperiodic points being isolated and then “the points in PrePer(fi) defined over our original field K coincide with the points in PrePer(fi) over Kbar” is confusing and, if K denotes the finitely generated field of definition, false (for f(x)=x^2 over Q, roots of unity are preperiodic but not defined over Q). The statement is not needed for the conclusion h_{f1} ≠ h_{f2}; please remove or rewrite it.
- [Proof of Theorem 1.3] The sentence “so we may assume that S contains no non-constant maps” should read “no constant maps”; as written it states the opposite of the intended reduction.
- [Theorem 1.2 statement] There is a typo: “endomorphsims” should be “endomorphisms”.
- [Proposition 2.1, Step 1] The maximal-prefix argument is compressed. The intended contradiction is that, if y ∈ α1(A) ∩ α2(A), then the second sequence pu2 shares a longer prefix with one of pα1v1 or pα2v2 depending on whether the first symbol of u2 after p is α1 or α2; spelling out these two cases would make the proof easier to follow.
- [Section 3, Heights] In the paragraph introducing heights, the sentence about positive characteristic and infinite fields should clarify that K is the finitely generated field of definition, not the original algebraically closed field, to avoid confusion with the global field K in the theorem statements.
- [Definition of Δ(S) and constants] The assertion that constant maps cannot be part of an independent pair and hence Δ(S)=Δ(S_{≥1}) deserves a brief justification, since a finite generating set may contain constants that are needed to generate other constant maps.
Circularity Check
No significant circularity: the derivation is self-contained; external theorems are imported as black boxes, not as fitted inputs or self-citations.
full rationale
The paper's chain of reasoning is Theorem 1.1 (a new ping-pong argument on the space of height functions) implying Theorem 1.2, and Theorem 1.2 combined with [BHPT24, Proposition 4.10] and [BG05] implying Theorem 1.3. No parameter is fitted to a subset of data and then called a prediction, and no definition is chosen so that the conclusion holds by construction. The central objects are preperiodic sets and canonical heights; the implication hf1 = hf2 iff PrePer(f1) = PrePer(f2) follows from the zero-locus property of canonical heights, which is independently cited. Lemma 3.1's use of [YZ21, Theorem 1.3] and [Car20, Theorem A] imports external rigidity theorems rather than assuming the target conclusion. Even if those theorems' exact hypotheses are not stated in the paper, that is a correctness or assumption gap, not circularity: the paper does not reduce its conclusion to its own inputs by construction, and it contains no self-citation chain that forces the result. The proof of Theorem 1.3 also relies on the independently established linear-case result of Breuillard and Gelander, which is external evidence rather than a restatement of the paper's conclusion. Therefore the honest finding is no significant circularity.
Assumptions & free parameters
assumptions (6)
- standard math A Weil or Moriwaki height h_L exists for an ample line bundle L on V over a finitely generated field, satisfying the Northcott property and functoriality for polarized endomorphisms.
- standard math Preperiodic points of a polarized endomorphism are Zariski dense, and the preperiodic points defined over the original algebraically closed field coincide with those over the algebraic closure of the finitely generated field.
- standard math If P is Zariski dense and P is contained in PrePer(w), then P equals PrePer(w) for an endomorphism w.
- domain assumption For a finitely generated subsemigroup S of End(P1) of non-polynomial growth with S≥2 nonempty, there exist f,g in S≥2 with PrePer(f) different from PrePer(g).
- standard math For any finitely generated field K and n at least 1 there is a constant c(n,K) such that every F in GL_n(K) generating a non-virtually-nilpotent group has Δ(F) at most c(n,K).
- standard math A finitely generated solvable group of polynomial growth is virtually nilpotent, so a non-polynomial-growth group is not virtually nilpotent.
Cite this review
Pith. "Pith review of A uniform Tits alternative for endomorphisms of the projective line." pith.science (2026). https://pith.science/paper/IQ6SJJZQ
@misc{pith2026250414263,
author = {Pith},
title = {Pith review of: A uniform Tits alternative for endomorphisms of the projective line},
year = {2026},
howpublished = {\url{https://pith.science/paper/IQ6SJJZQ}},
note = {Machine review of arXiv:2504.14263}
}
abstract
A recent article of J.P. Bell, K. Huang, W. Peng and T.J. Tucker establishes an analog of the Tits alternative for semigroups of endomorphisms of the projective line. The proof involves a ping-pong argument on arithmetic height functions. Extending this method, we obtain a uniform version of the same alternative. In particular, we show that semigroups of $\mathrm{End}(\mathbb{P}^{1})$ of exponential growth are of uniform exponential growth.
Reference graph
Works this paper leans on
-
[1]
E. Breuillard and T. Gelander, Cheeger constant and algebraic entropy of linear groups, Int. Math. Res. Not. 2005 (2005), no. 56, 3511--3523
work page 2005
-
[2]
E. Bombieri and W. Gubler, Heights in diophantine geometry, Cambridge University Press, Cambridge, 2006
work page 2006
-
[3]
M. F. Barnsley and A. N. Harrington, Mandelbrot set for pairs of linear maps, Physica D: Nonlinear Phenomena 15 (1985), no. 3, 421--432
work page 1985
-
[4]
J. P. Bell, K. Huang, W. Peng, and T. J. Tucker, A tits alternative for endormophisms of the projective line, J. Eur. Math. Soc. 26 (2024), no. 12, 4903--4922
work page 2024
-
[5]
On uniform exponential growth for solvable groups
E. Breuillard, On uniform exponential growth for solvable groups, 2006, arXiv preprint math/0602076
work page Pith review arXiv 2006
-
[6]
Heights and arithmetic dynamics over finitely generated fields
A. Carney, Heights and arithmetic dynamics over finitely generated fields, 2020, arXiv preprint arXiv:2010.07200
work page Pith review arXiv 2020
-
[7]
Fakhruddin, Questions on self maps of algebraic varieties, J
N. Fakhruddin, Questions on self maps of algebraic varieties, J. Ramanujan Math. Soc. 18 (2003), no. 2, 109--122
work page 2003
-
[8]
Falconer, The geometry of fractal sets, Cambridge university press, Cambridge, 1985
K. Falconer, The geometry of fractal sets, Cambridge university press, Cambridge, 1985
work page 1985
Show all 14 references
-
[9]
Moriwaki, Arithmetic height functions over finitely generated fields, Invent
A. Moriwaki, Arithmetic height functions over finitely generated fields, Invent. Math. 140 (2000), no. 1, 101--141
2000
-
[10]
Okniński, Semigroups of matrices, World Scientific, 1998
J. Okniński, Semigroups of matrices, World Scientific, 1998
1998
-
[11]
Serre, Analogues kählériens de certaines conjectures de weil, Ann
J.-P. Serre, Analogues kählériens de certaines conjectures de weil, Ann. Math. 71 (1960), no. 2, 392--394
1960
-
[12]
J. A. Wolf, Growth of finitely generated solvable groups and curvature of riemannian manifolds, J. Differ. Geom. 2 (1968), 421--446
1968
-
[13]
Yuan and S
X. Yuan and S. Zhang, The arithmetic hodge index theorem for adelic line bundles ii: finitely generated fields, 2021, arXiv preprint arXiv:1304.3539v2
2021 arXiv
-
[14]
Zhang, Small points and adelic metrics, J
S. Zhang, Small points and adelic metrics, J. Algebr. Geom. 4 (1995), no. 2, 281--300
1995
Reviewed August 16, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.