{"id":"38283950-5d98-43ae-8881-69506c636679","arxiv_id":"1908.03485","paper_version":2,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Isogenous Drinfeld modules have graded heights differing by at most deg N plus an explicit constant, with an explicit coefficient bound for rank-2 Drinfeld modular polynomials.","lead":"This paper proves explicit bounds on how much the height of a Drinfeld module can change under an isogeny, and gives an explicit upper bound on the size of the coefficients of Drinfeld modular polynomials in rank 2. It is a function-field analogue of known elliptic-curve results, using Taguchi's and Gekeler's machinery.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.1(1)'s explicit constant rests on unverified Gekeler estimates in Lemma 5.3; a small-case check should confirm the monotonicity and boundary-value claims.","rationale":"The paper's main contribution is the explicit height-difference bound; its proof goes through Lemma 5.3, which imports the crucial analytic estimates from Gekeler. The reader identified this as the weakest assumption, and I agree. The paper does not state the precise forms of the cited results, and the interpolation argument in Lemma 5.3 requires that log|g_i| be constant on fibres and monotone away from the origin on the entire building; a misstatement would invalidate the bound. A separate internal error appears in the proof of Corollary 3.3: the claimed inequality h(φ) ≤ (q^r−1)h_G(φ) is false because h(φ) depends on the representative, e.g., scaling g_i by c^{q^i−1} multiplies h(g_i) by (q^i−1)h(c) while h_G is invariant. This does not affect Theorem 3.1 but does invalidate the proof of the finiteness corollary as written. I recommend a conditional acceptance: the authors should verify Lemma 5.3 against Gekeler's exact statements and repair the inequality.","tokens_in":13228,"tokens_out":57878,"duration_ms":558876,"concrete_test":"Verify the precise statements of [Gek17, Prop. 4.10, Cor. 4.11, Cor. 4.16, §4.6] and check whether they imply the monotonicity of log|g_i| on all of W(Q) and the equalities (21)–(22) on the full strata. In parallel, compute the coefficients for the rank-2 lattices Λ_k = A + A t^k over F_2 and F_3 (k=0,1,2) from the exponential expansion, and verify Lemma 5.3's bounds: log|g_1| = q for k≥1, log|g_2| ≤ q^2 (equality at k=0), and max(log|g_1|, (1/(q^2−1))log|g_2|) ∈ [q^2/(q^2−1), q/(q−1)]. Any violation falsifies the lemma and the constant in Theorem 3.1(1).","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Lemma 5.3 is the linchpin of the proof: its upper bound q/(q−1) enters Theorem 3.1(1) via (23), and its lower bound q^r/(q^r−1) gives the negative part of the constant. The proof imports from [Gek17] three statements: (i) at the origin of W(Q), log|g_r(ω)| = q^r and log|g_i(ω)| ≤ q^i for i<r, with equality attained on λ^{-1}(o); (ii) each log|g_i(ω)| is non-increasing as λ(ω) moves away from o in the building; (iii) on the strata F \\ F_{r−1}, (F_{r−1}∩...∩F_{r−i+1})\\F_{r−i}, the equalities (21)–(22) hold. The paper does not quote the precise statements of Cor. 4.11/4.16, nor does it address boundary cases such as zeros of g_i. If (ii) is only a maximum principle on a proper subset, the upper bound can be violated elsewhere; if (iii) fails because Cor. 4.16 gives values only at vertices rather than on full fibres, the lower bound breaks. Since both the explicit height bound and the rank-2 modular polynomial bound rest on this lemma, this is the most load-bearing point. A separate flaw: the proof of Cor. 3.3 uses h(φ) ≤ (q^r−1)h_G(φ), which is false for arbitrary representatives.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves explicit bounds on the change of the graded height h_G of a Drinfeld module under an isogeny. The main result, Theorem 3.1(1), states that for an isogeny f: φ → φ' of rank r Drinfeld modules with ker f ⊂ φ[N], one has |h_G(φ') − h_G(φ)| ≤ deg N + (q/(q−1) − q^r/(q^r−1)). The proof combines Taguchi's isogeny lemma, the equality of the finite part of h_G with Taguchi's height, and analytic estimates on reduced lattices taken from Gekeler's work on Drinfeld modular forms and Bruhat–Tits buildings. In the rank 2 case the authors also prove a one-sided inequality involving the usual j-invariant and log deg f. These results are applied to obtain an effective finiteness statement for isogeny classes and, via an interpolation argument, an explicit upper bound on the height of Drinfeld modular polynomials Φ_m, complementing Hsia's asymptotic.","tokens_in":13532,"tokens_out":16455,"duration_ms":172968,"significance":"If the main theorem holds, it provides the first fully explicit function-field analogue of Pazuki's elliptic-curve height comparison, with the constant in Theorem 3.1(1) depending only on q and r. The paper is well structured: the heavy inputs (Taguchi's isogeny lemma, Gekeler's building estimates, David–Denis's effective isogeny theorem, Hsia's asymptotic) are clearly cited, and the authors are transparent about the analytic estimates that drive the explicit constant. The derivation of the modular polynomial bound from the height inequality is elegant and gives a concrete statement complementing Hsia's asymptotic. However, two points need attention: one inequality in the proof of Corollary 3.3 is false as stated, and Lemma 5.3 relies on building-theoretic statements that are not quoted precisely enough for the reader to verify the constant. These are load-bearing for the finiteness application and for the explicit constant, respectively.","major_comments":[{"comment":"The proof uses the inequality h(φ) ≤ (q^r − 1)h_G(φ), where h(φ) = max_i h(g_i) is the maximum of the Weil heights of the coefficients. This inequality is false for arbitrary representatives. For example, take q = 2, r = 2, and φ_t(X) = tX + t^N X^2 + t^{3N} X^4 with N ≥ 1. Then h(φ) = 3N, while j_1 = g_1^3/g_2 = 1 and j_2 = 1, so h_G(φ) = 0. More generally, h_G is invariant under isomorphism, whereas max_i h(g_i) is not, so no such inequality can hold without fixing a normalized representative. Since the proof then uses [DD99, Thm. 1.3] to bound deg f in terms of h(φ) and converts the result into a bound in terms of h_G(φ), the derivation of Corollary 3.3 does not work as written. This also affects the new proof of Corollary 3.4. Please repair this step, for example by applying [DD99] to a height that is genuinely comparable to h_G, or by proving the needed comparison after choosing a suitable representative.","section":"§3, proof of Corollary 3.3"},{"comment":"The explicit constant in Theorem 3.1(1) rests entirely on Lemma 5.3, whose proof imports from [Gek17, Cor. 4.11 and 4.16] the assertions that each log|g_i(ω)| is non-increasing as λ(ω) moves away from the origin and that equations (21)–(22) hold on the stated strata. The paper does not quote the precise statements of those corollaries, nor does it address the boundary cases (for instance, fibres over boundary points of W(Q) or possible zeros of g_i). Since (19) enters both the bound on part (B) and the constant in Theorem 3.1, and since Theorem 3.1(2) and Proposition 6.5 inherit this constant, the proof would be substantially easier to certify if the exact statements from [Gek17] were reproduced or if a direct verification of the monotonicity and fibre-constancy claims were supplied. I am not asserting that the cited estimates are wrong, but as written Lemma 5.3 is a verification gap in the central argument.","section":"§5.1, Lemma 5.3"}],"minor_comments":[{"comment":"In the proof of Lemma 6.3, the sentence 'by Lemma 6.2 we have h(T_k) ≤ q^{nd}' should read h(T_k) ≤ nd (or h(T_k) = log max |a_j| ≤ nd). The subsequent '+2nd' in (24) only makes sense with the logarithmic bound.","section":"§6, Lemma 6.3"},{"comment":"The notation n_σ in (17) is used for the local degree of the embedding σ; it is not defined in the text. Please define n_σ explicitly when the sum over embeddings is introduced.","section":"§5.1, equation (17)"},{"comment":"Please clarify the normalization of the functions g_i(ω) in the building-coordinate argument: are these the coefficients of the Drinfeld module associated to the lattice generated by ω, and are they evaluated after fixing the basis (ω_1,...,ω_r) with ω_r = 1? The notation log|g_i(ω)| is otherwise ambiguous, since the coefficients depend on the lattice, not on the chosen basis.","section":"§5.1, Lemma 5.3"},{"comment":"The displayed bound for h(Φ_m) is hard to parse because of the long bracketed expression. Adding a line break or a named auxiliary quantity (for example writing the max argument as B) would improve readability.","section":"§6, Proposition 6.5"}],"recommendation":"major_revision","confidential_remarks":"The main height inequality (Theorem 3.1) appears plausible and is well motivated, but the proof of Corollary 3.3 contains a definite false inequality, and Lemma 5.3 leaves a verification gap around the Gekeler estimates. These are fixable within the scope of the paper, so I recommend major revision rather than rejection. The referee should probably consult [Gek17] directly to certify Lemma 5.3 before acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nHere's my read on Breuer, Pazuki, and Razafinjatovo's paper. The real contribution is the pair of explicit inequalities in Theorem 3.1: the graded-height difference under an isogeny is bounded by deg N plus an explicit constant depending only on rank and field size, and the rank-2 variant in terms of j-invariants. These are genuinely new, quantitative analogues of Pazuki's elliptic curve result. The strategy is transparent: decompose the height difference into a Taguchi-height term, an analytic term, and a lattice term, and bound each piece with existing machinery (Taguchi's isogeny lemma, Gekeler's building estimates, and a new covolume lemma). The modular polynomial coefficient bound in Proposition 6.5 is also new and follows cleanly from the rank-2 inequality.\n\nThere is a real flaw in the proof of Corollary 3.3. The authors claim 'it is easy to see' that h(φ) ≤ (q^r−1)h_G(φ), where h(φ)=max_i h(g_i). This is false. For example, over F_2(t), take g1=t^{-N}, g2=t^{-3N} with r=2. Then h_G=0 (the local contributions cancel) while h(g2)=3N, so the inequality fails by an arbitrarily large factor. The proof of Corollary 3.3 collapses at that line. The corollary itself may be salvageable, since the finiteness result is already due to Taguchi, but the quantitative bound as stated needs a different argument.\n\nThe other soft spot is Lemma 5.3, the linchpin for the explicit constant in Theorem 3.1(1). The proof imports several claims from Gekeler's Corollaries 4.11 and 4.16: monotonicity of log|g_i| along the building, boundary values on strata, and equality at the origin. The paper doesn't quote the actual statements or address boundary cases. This may be fine, but a referee should ask the authors to state the precise Gekeler results or add a verification for small ranks. The constant depends entirely on this lemma.\n\nThe interpolation lemma in Section 6 has minor notational slips, but they don't affect the argument. The self-citation to [Paz19] is appropriate as a template rather than a premise.\n\nOverall, the central theorems look sound and the quantitative bounds are of real interest to people working on Drinfeld modules. The Corollary 3.3 flaw is a genuine gap, but it's a secondary corollary and likely fixable. I'd send this to a serious referee, with instructions to ask for a corrected proof or a softened statement for Corollary 3.3, and to clarify the Gekeler input in Lemma 5.3. Not a desk reject.","headline":"Main explicit height bounds are new and plausible, but the proof of Corollary 3.3 contains a false inequality and Lemma 5.3 leans on unstated Gekeler estimates, so this deserves a careful revision rather than rejection.","tokens_in":14081,"tokens_out":13900,"would_cite":true,"duration_ms":122459,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["11G09","11G50","14G17","14G40"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves explicit bounds on how much an isogeny can change the height of a Drinfeld module, and applies them to finiteness and modular polynomials.","keywords":["Drinfeld modules","graded height","isogenies","Taguchi height","modular polynomials","function fields","Bruhat-Tits building","finiteness of isogeny classes"],"falsifier":"Find a reduced Drinfeld module of rank $r$ over $\\mathbb{C}_\\infty$ for which $\\log \\max_i |g_i|^{1/(q^i-1)}$ exceeds $\\frac{q}{q-1}$ or falls below $\\frac{q^r}{q^r-1}$; either example would refute Lemma 5.3 and with it the explicit constant in Theorem 3.1.","tokens_in":13030,"feed_emoji":"📐","tokens_out":10570,"duration_ms":104522,"temperature":0.7,"pith_summary":"The paper proves a function-field analogue of the classical elliptic-curve statement that an isogeny changes the Faltings height by at most a small multiple of the logarithm of the degree. For Drinfeld $\\mathbb{F}_q[t]$-modules of rank $r$ in generic characteristic, it shows that if $f:\\varphi\\to\\varphi'$ is an isogeny with $\\ker f \\subset \\varphi[N]$, then $|h_G(\\varphi')-h_G(\\varphi)|\\leq \\deg N+\\left(\\frac{q}{q-1}-\\frac{q^r}{q^r-1}\\right)$, where $h_G$ is the graded height computed from the coefficients of the module. The additive constant depends only on $q$ and $r$, so the height difference is controlled by the size of the kernel up to a rank-uniform constant. This comparison is then used to give an effective finiteness theorem for isogeny classes and, in rank two, an explicit upper bound on the coefficients of Drinfeld modular polynomials.","feed_headline":"Explicit bound: isogenous Drinfeld modules have close heights","feed_subtitle":"It also makes isogeny classes effectively finite and bounds the size of modular polynomial coefficients.","key_machinery":"The argument decomposes the height difference into a finite-place part governed by the isogeny lemma for the differential height and an infinite-place part that is handled by reducing each Drinfeld module to a reduced one and estimating the coefficient sizes of reduced modules. The load-bearing object is the fundamental domain for the moduli space and its associated Bruhat-Tits building, the simplicial complex encoding the relative sizes of the basis vectors of a lattice. The paper's Lemma 5.3 uses the building structure to show that $\\log \\max_i |g_i|^{1/(q^i-1)}$ lies between $q^r/(q^r-1)$ and $q/(q-1)$ for every reduced rank-$r$ module, and that two-sided bound converts the infinite-place term into the explicit additive constant in Theorem 3.1.","core_discovery":"The paper's central claim is that the graded height $h_G$, the logarithmic size of the coefficient tuple viewed in weighted projective space, changes very little under isogeny. For an isogeny $f:\\varphi\\to\\varphi'$ of rank $r$ modules with $\\ker f \\subset \\varphi[N]$, Theorem 3.1 gives $$|h_G(\\varphi')-h_G(\\varphi)|\\leq \\deg N+\\left(\\frac{q}{q-1}-\\frac{q^r}{q^r-1}\\right),$$ where $N$ is the element of $\\mathbb{F}_q[t]$ of minimal degree whose $N$-torsion contains the kernel. In rank two this yields the refined $j$-invariant inequality $$h(j')-h(j)\\leq \\frac{$q^{2}$-1}{2}\\log\\deg f+\\frac{$q^{2}$-1}{2}\\log\\left(1+\\frac{h(j')}{q}\\right)+q,$$ with the same shape as the elliptic-curve comparison. From these bounds the paper derives an effective finiteness theorem for isogeny classes and an explicit upper bound on the coefficient height of Drinfeld modular polynomials.","pith_inferences":["Although the paper does not discuss optimality, if the coefficient bound in Lemma 5.3 is tight then the additive constant in Theorem 3.1 cannot be improved for generic kernels.","The same decomposition may apply to isogenies between Drinfeld modules of different ranks, or to heights of higher-rank modular polynomials, since the only rank-specific input is the two-sided coefficient bound for reduced modules.","A direct computation of reduced rank-3 modules with lattices at different vertices of the fundamental domain could test whether the extremes in Lemma 5.3 are actually attained, which would tell whether the constant in Theorem 3.1 is sharp."],"forward_implications":["For a Drinfeld module defined over a finite extension of $\\mathbb{F}_q(t)$, the height bound combines with the standard finiteness property of heights to give an effective version of the finiteness theorem: each isogeny class contains finitely many isomorphism classes, with a computable bound on their number.","For rank 2, the inequality for the $j$-invariant yields an explicit upper bound on the height $h(\\Phi_m)$ of the coefficients of the modular polynomial $\\Phi_m(X,Y)$; the paper records the asymptotic form $h(\\Phi_m)<\\left(\\frac{q^2+4}{2}+\\varepsilon\\right)\\psi(m)\\deg(m)$, close to the known exact asymptotics for these polynomials.","The proof gives a template for controlling the difference between any two height functions that agree at finite places and differ only through infinite-place symmetries, whenever the moduli space has a fundamental domain with explicit bounds on the coordinate functions.","The theorem's rank-2 statement has the same logarithmic shape as the classical elliptic-curve comparison, indicating that the function-field analogue holds with a fully explicit constant rather than an asymptotic one."],"supporting_citations":[{"why":"Defines the differential height and proves the isogeny lemma used to bound the finite-place part of the height difference.","marker":"[Tag93]"},{"why":"Supplies the building-theoretic estimates on coefficient sizes of reduced Drinfeld modules used as Lemma 5.3.","marker":"[Gek17]"},{"why":"Supplies the estimate of the $j$-invariant of a reduced rank-2 lattice used in Lemma 5.4.","marker":"[Gek97]"},{"why":"Provides the minimal-isogeny and stable-reduction results used for the effective finiteness corollary.","marker":"[DD99]"},{"why":"Gives the asymptotic height result for modular polynomials that the explicit bound in Proposition 6.5 is compared with.","marker":"[Hsi98]"},{"why":"Constructs the Drinfeld modular polynomials whose coefficient heights Proposition 6.5 bounds.","marker":"[Bae92]"},{"why":"States the elliptic-curve height comparison whose argument is adapted to Drinfeld modules.","marker":"[Paz19]"}],"fun_headline_variants":["Explicit height bounds for isogenous Drinfeld modules","Isogeny classes of Drinfeld modules: effective finiteness","Drinfeld isogenies: height gaps and finite classes","Height difference under Drinfeld isogeny: explicit bound","Rank-2 Drinfeld isogenies: modular polynomial size bound"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof assumes that a cited estimate pinning the coefficient sizes of reduced Drinfeld modules between two explicit bounds holds without extra hypotheses; if that estimate is wrong or incomplete, the additive constant in Theorem 3.1 does not follow.","fun_headline_variants_meta":{"raw":{"variants":["Explicit height bounds for isogenous Drinfeld modules","Isogeny classes of Drinfeld modules: effective finiteness","Drinfeld isogenies: height gaps and finite classes","Height difference under Drinfeld isogeny: explicit bound","Rank-2 Drinfeld isogenies: modular polynomial size bound"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000616,"raw_usage":{"total_tokens":2795,"prompt_tokens":813,"completion_tokens":1982,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":429,"completion_tokens_details":{"reasoning_tokens":1894}},"tokens_in":429,"tokens_out":1982,"duration_ms":15606,"temperature":1.0,"reasoning_tokens":1894,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:11:25.030501+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a reduced Drinfeld module of rank $r$ over $\\mathbb{C}_\\infty$ for which $\\log \\max_i |g_i|^{1/(q^i-1)}$ exceeds $\\frac{q}{q-1}$ or falls below $\\frac{q^r}{q^r-1}$; either example would refute Lemma 5.3 and with it the explicit constant in Theorem 3.1.","supporting_citations":[],"review_version":1}