{"id":"59e10c47-ef03-4def-9875-c497f71f24a5","arxiv_id":"2504.14263","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Finitely generated semigroups of endomorphisms of P1 over characteristic 0 that have exponential growth have uniform exponential growth, with diameter of independence at most 2 when a degree at least 2 map exists.","lead":"A short arithmetic-dynamics note proves a uniform version of the Tits alternative for semigroups of algebraic self-maps of the projective line. If a semigroup has exponential word growth, it has uniform exponential growth, meaning the growth rate is bounded below by a constant independent of the generating set.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central Lemma 3.1 depends on an unstated arithmetic-dynamical rigidity upgrade from Zariski-dense inclusion to equality of preperiodic sets; if [YZ21, Thm 1.3] or [Car20, Thm A] does not have exactly this implication, the proof of Theorems 1.2–1.3 collapses.","rationale":"I read the paper as a short, well-organized proof that refines the height-based ping-pong of BHPT24 into a uniform statement. The contraction lemma in Section 2 appears correct: the c1 + c2 ≤ 1 condition is used genuinely, the connected-attractor Hausdorff-measure argument is standard, and the sharpness remark is coherent. The construction of canonical heights on H_L and the proof of Theorem 1.1 are also sound. The proof of Theorem 1.3 splits into a degree-1 linear case, handled by the cited Breuillard–Gelander bound after the implicit PGL2-to-GL3 embedding, and a S_{\\ge 2} case handled by Theorem 1.2. Theorem 1.2 depends on Lemma 3.1. Within Lemma 3.1, the statement 'by the Northcott property, all w-orbits in P are finite' is not justified as written, but it is repairable by the canonical-height comparison sketched above, so I do not treat it as the central objection. The central objection is the external rigidity upgrade from Zariski-dense inclusion P ⊂ PrePer(w) to equality P = PrePer(w). The paper cites [YZ21, Theorem 1.3, (3) ⇒ (1)] and [Car20, Theorem A, (3) ⇒ (4)] without stating their hypotheses. The reader identified this same assumption, and I agree. Because the entire uniform bound for S_{\\ge 2} relies on this upgrade, and because a misreading of the cited theorem would be circular (if the theorem requires PrePer(f) = PrePer(g) as an input), acceptance should be conditional on verification of the cited theorem's exact statement and applicability. I am not aware of evidence that the theorem is wrong; the concern is verification, not known falsity. Hence CONDITIONAL rather than REJECT or UNCHANGED.","tokens_in":6903,"tokens_out":28407,"duration_ms":255791,"concrete_test":"Retrieve the exact statements of [YZ21, Theorem 1.3] and [Car20, Theorem A]. Verify that in each theorem the hypothesis labelled (3) is precisely 'PrePer(f) contains a Zariski dense subset of PrePer(g)' (or equivalent) for polarized endomorphisms f, g of a projective variety over a finitely generated field, and that the implication (3) ⇒ (1) concludes h_f = h_g (from which PrePer(f) = PrePer(g) follows by the Northcott property). Check that all conditions are met in Lemma 3.1: f has algebraic degree at least 2, w is semi-polarized by the same ample line bundle, the field is finitely generated, and P = PrePer(f) is Zariski dense by [Fak03, Theorem 5.1]. If the theorem instead uses equality of preperiodic sets as a hypothesis, Lemma 3.1 is circular and the proof of Theorem 1.3 fails.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Theorem 1.3 is built on Theorem 1.2, whose proof uses Lemma 3.1. In Lemma 3.1, after proving that P = PrePer(f) is invariant under every generator (so w(P) = P), the text says 'by the Northcott property, all w-orbits in P are finite' to conclude P ⊂ PrePer(w). This is not immediate from Northcott because the degrees [K(w^n(x)) : K] grow with n. The step is repairable by comparing canonical heights: h_f(w^n(x)) = 0, h_f − h_w is bounded, and h_w(w^n(x)) = d_w^n h_w(x), so d_w^n h_w(x) is bounded and h_w(x) = 0, whence x ∈ PrePer(w). Thus the real load-bearing step is the cited upgrade: 'apply [YZ21, Theorem 1.3, (3) ⇒ (1)], or [Car20, Theorem A, (3) ⇒ (4)] ... to conclude that P = PrePer(w).' The paper states no hypotheses of these theorems. The argument needs exactly the implication: if P is Zariski dense and contained in PrePer(w), then PrePer(f) = PrePer(w). If the cited theorem instead has PrePer(f) = PrePer(g) as a hypothesis rather than a conclusion, the argument is circular; if it requires extra conditions not met here (for example, a common iterate, or both maps defined over a number field, or degree strictly greater than 2), Lemma 3.1 fails. Since Lemma 3.1 supplies the distinct-preperiodic-set pair needed to invoke Theorem 1.1 in Theorem 1.2, and Theorem 1.3 uses Theorem 1.2 for the S_{\\ge 2} case, the central claim rests on this external rigidity theorem.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","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.","tokens_in":7282,"tokens_out":26735,"duration_ms":256157,"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":[{"comment":"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":"Section 3, Lemma 3.1"},{"comment":"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.","section":"Section 3, Lemma 3.1"}],"minor_comments":[{"comment":"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.","section":"Proof of Theorem 1.1"},{"comment":"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.","section":"Proof of Theorem 1.3"},{"comment":"There is a typo: “endomorphsims” should be “endomorphisms”.","section":"Theorem 1.2 statement"},{"comment":"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":"Proposition 2.1, Step 1"},{"comment":"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.","section":"Section 3, Heights"},{"comment":"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.","section":"Definition of Δ(S) and constants"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid and interesting extension of BHPT24, and the height ping-pong itself appears sound. My main concern is that Lemma 3.1 relies on a cited rigidity theorem whose exact statement is not given; I did not find the theorem text in the arXiv version of YZ21 that is cited, so I could not fully verify the implication being used. Asking the authors to state the theorem and verify its hypotheses is the right request. The Northcott step that a skeptic might flag is actually fine, once the one-line degree-bound argument is added."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper gives the uniform version of the BHPT24 Tits alternative: finitely generated subsemigroups of End(P1) over characteristic zero are either of polynomial growth or have finite diameter of independence, hence uniform exponential growth. That answers Question 5.1 of BHPT24. Theorem 1.1 is also new: distinct preperiodic sets for two polarized endomorphisms of the same projective variety force a free semigroup. Both results are real, and the technique is worth knowing.\n\nThe main new ingredient is Proposition 2.1, a contraction ping-pong lemma on metric spaces with ratio sum c1+c2 ≤ 1, proved using Hausdorff measure in the connected case. The proof is elegant and apparently correct. The application to the space of bounded height functions is natural and gives the uniform bound Δ(S) ≤ 2 in the non-linear case. The argument importing [BHPT24, Prop 4.10] and [BG05] for the linear case is standard.\n\nThe soft spots are in Lemma 3.1. The line \"by the Northcott property, all w-orbits in P are finite\" is too fast: Northcott alone does not control degree growth along w-orbits. The stress-test note is right about this. But the conclusion P ⊂ PrePer(w) is repairable by comparing h_f and h_w, so this is an expository gap, not a mathematical one. The genuine issue is the next step: the paper invokes [YZ21, Theorem 1.3, (3)⇒(1)] or [Car20, Theorem A, (3)⇒(4)] to upgrade P ⊂ PrePer(w) to equality, without stating what those theorems assume or conclude. A referee should verify that the quoted implication has exactly the needed hypotheses: Zariski-dense invariant subset of PrePer(w), with no extra condition such as a common iterate or number-field restriction. If those hypotheses are not met, Lemma 3.1 fails, and Theorems 1.2–1.3 would collapse. I have no reason to think the citation is wrong, but it is load-bearing enough that the paper should spell it out.\n\nMinor stuff: a few typos (\"endomorphsims\", \"non-emtpy\") and the maximal-prefix argument in Step 1 of Proposition 2.1 is compressed. None of this affects the mathematics.\n\nWho this is for: people working in arithmetic dynamics, growth of semigroups, and height theory. The paper is short, clear, and deserves a serious referee. I would send it out, asking the referee to check the YZ21/Car20 quote and to have the author clarify the Northcott step.","headline":"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.","tokens_in":7833,"tokens_out":6147,"would_cite":true,"duration_ms":53485,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37P05","37P30"],"pacs":[],"model":"deepseek-v4-flash","headline":"Two polarized endomorphisms with distinct preperiodic sets generate a free semigroup, and exponential-growth subsemigroups of End(P1) have uniform exponential growth.","keywords":["Tits alternative","uniform exponential growth","preperiodic points","canonical heights","ping-pong lemma","endomorphisms of the projective line","algebraic entropy","arithmetic dynamics"],"falsifier":"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.","tokens_in":6671,"feed_emoji":"📈","tokens_out":16308,"duration_ms":133753,"temperature":0.7,"pith_summary":"This paper proves a uniform version of the Tits alternative for semigroups of algebraic self-maps of the projective line. The central claim is that two endomorphisms polarized by the same line bundle and having different sets of preperiodic points must generate a free semigroup of rank two. From this, any finitely generated subsemigroup of $\\mathrm{End}(\\mathbb{P}^1)$ over a field of characteristic zero is either of polynomial growth or has a finite diameter of independence—meaning there is a fixed word length within which two independent elements appear; in particular, semigroups of exponential growth grow uniformly exponentially, with algebraic entropy at least $\\log(2)/2$ when a degree-at-least-two map is present. The proof sends each endomorphism to a contraction on a space of height functions whose fixed point is the canonical height, then applies a ping-pong lemma for contractions with distinct fixed points. This answers a question left open in earlier work and makes the polynomial-versus-exponential dichotomy quantitative.","feed_headline":"Distinct preperiodic sets force a free semigroup","feed_subtitle":"A quantitative Tits alternative: exponential-growth semigroups of self-maps of P1 grow at a definite rate.","key_machinery":"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 $\\Sigma(F) \\geq \\log(2)/\\Delta(F)$.","core_discovery":"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.","pith_inferences":["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."],"forward_implications":["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."],"supporting_citations":[{"why":"Establishes the non-uniform Tits alternative for End(P1), supplies the height ping-pong method and the result used to produce two elements with distinct preperiodic sets, and poses the question answered here.","marker":"[BHPT24]"},{"why":"Provides the attractor self-similarity theorem and the Hausdorff measure facts used in Proposition 2.1.","marker":"[Fal85]"},{"why":"The disconnected-attractor step of Proposition 2.1 is adapted from this source.","marker":"[BH85]"},{"why":"Supplies the Weil height construction and the Northcott and functoriality properties used to set up the height-function space.","marker":"[BG06]"},{"why":"Defines Moriwaki heights over finitely generated fields in characteristic zero, the height notion used in the main proof.","marker":"[Mor00]"},{"why":"Gives the finite bound on the diameter of independence for non-virtually-nilpotent linear groups, handling the degree-one case of Theorem 1.3.","marker":"[BG05]"},{"why":"Shows that non-polynomial-growth linear groups are not virtually nilpotent, which triggers the application of [BG05].","marker":"[Wol68]"},{"why":"The arithmetic rigidity theorem used in Lemma 3.1 to upgrade inclusion of a Zariski-dense preperiodic set to equality.","marker":"[YZ21]"},{"why":"Supplies the same rigidity upgrade in positive characteristic, covering cases where [YZ21] does not apply.","marker":"[Car20]"},{"why":"Gives Zariski density of preperiodic sets and isolation of preperiodic points, used in Theorem 1.1 and Lemma 3.1.","marker":"[Fak03]"}],"fun_headline_variants":["Uniform Tits alternative: exponential growth has a definite rate","Distinct preperiodic sets force a free semigroup","Ping-pong on heights: uniform growth for P1 semigroups","Exponential growth in End(P1) is always uniform","Preperiodic sets decide freeness; growth rate becomes uniform"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"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.","fun_headline_variants_meta":{"raw":{"variants":["Uniform Tits alternative: exponential growth has a definite rate","Distinct preperiodic sets force a free semigroup","Ping-pong on heights: uniform growth for P1 semigroups","Exponential growth in End(P1) is always uniform","Preperiodic sets decide freeness; growth rate becomes uniform"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000773,"raw_usage":{"total_tokens":3384,"prompt_tokens":872,"completion_tokens":2512,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":488,"completion_tokens_details":{"reasoning_tokens":2426}},"tokens_in":488,"tokens_out":2512,"duration_ms":15559,"temperature":1.0,"reasoning_tokens":2426,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T11:54:24.376083+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"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.","supporting_citations":[],"review_version":1}