{"id":"f3f639fc-4dbc-4295-9feb-02cc98a21c69","arxiv_id":"2607.09565","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.5,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Explicit constants are obtained for Bombieri–André G-function height bounds and applied to give polynomial-in-degree height bounds for lines in tori intersecting codimension-2 subgroups.","lead":"The paper derives fully explicit constants in Bombieri–André height bounds for unexpected global polynomial relations among G-function values. It then applies them to produce concrete height bounds for unlikely intersections of lines through the identity with codimension-2 subgroups in algebraic tori, improving Habegger’s bound for points of small Galois degree.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified","rationale":"The paper’s strongest claim is an explicit height bound under precisely the hypotheses needed for the G-function method to produce Q-coefficient relations of degree controlled by τ(K(s)). Every step from the corrected local inequality through the Hilbert-function optimisation to the torus application is written out with explicit intermediate bounds; the only non-formal ingredients (Ax–Schanuel, known Hilbert-function estimates) are standard and correctly invoked. The reader’s weakest_assumption accurately describes why the method cannot reach Σ_1 or curves missing torsion points, but that is a limitation of scope, not a load-bearing error inside the theorems that are proved. Consequently no adjustment to the ACCEPT verdict is warranted. The suggested numerical check simply confirms that the final arithmetic of the constants has been performed correctly.","tokens_in":29718,"tokens_out":457,"duration_ms":4746,"concrete_test":"Independently recompute the numerical constants of Corollary 1.3 by substituting n=3, λ=4 (δ=1) into the expanded form (21) of Theorem 4.2 and n=3, c6(3)=7/2, c7(3)=9/2 into (16) of Corollary 5.5; verify that the resulting coefficients are exactly 2708 and 166 as claimed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (Theorem 6.1) rests on a fully explicit, self-contained derivation: corrected André inequality (Prop. 4.6), construction of (r_v)-global relations of controlled degree via multiplicative characters (Prop. 6.8), Ax–Schanuel independence (Lemma 6.2), and concrete input data for radii/sizes (Lemmas 6.9–6.12). The geometric restrictions (line through identity, Σ_2 only) are essential to the method and openly stated; they do not create an internal gap in the argument as written. No free parameters, circularity, or unverified asymptotic steps appear. The reader’s weakest_assumption correctly flags the method’s scope limits rather than a flaw in the proof of the stated theorems.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The paper gives fully explicit constants in Bombieri–André height bounds for points at which G-functions satisfy non-trivial (or super-strongly non-trivial) global polynomial relations. After correcting minor errors in André’s inequality (Proposition 4.6), it proves refined bounds (Theorems 4.2, 5.1–5.2 and Corollaries 4.4–4.5, 5.4–5.5) that track the number of independent relations, the Hilbert function of the ideal of functional relations, and an acceptable system of radii (r_v). These are applied to lines C through the identity in G_m^n that are not contained in a proper algebraic subgroup: for s in C(K) ∩ Σ_2 one obtains an explicit height bound for the local parameter x(s) that is polynomial in δ = max{1, τ(K(s))} and linear in a height H of the coefficients of the line (Theorem 6.1 and Corollary 1.3). The input data (singularities, radii, sizes of the logarithmic G-functions and of the associated differential operator) are computed in Lemmas 6.9–6.12, and the (r_v)-global relations of controlled degree are constructed in Proposition 6.8 via multiplicative characters.","tokens_in":29878,"tokens_out":867,"duration_ms":7406,"significance":"Explicit constants for the Bombieri–André principle have been missing for decades; the paper supplies them in several usable forms and demonstrates that the G-function method yields competitive height bounds for low-degree points on lines in tori (better than Habegger’s bound for [K(s):Q] up to roughly 10^4 even after optimising Habegger for lines). The introduction of (r_v)-global relations cleanly handles the usual radius mismatch and produces sharper constants. The application is the first written instance of the method for algebraic groups rather than Shimura varieties, and the inexplicit existence argument in §6.A is a useful pedagogical illustration. All constants are derived from first principles with no free parameters; the geometric restrictions (line through a torsion point, Σ_2 only) are essential to the method and are stated openly.","major_comments":[],"minor_comments":[{"comment":"Abstract and title: “polymomial” should be “polynomial”.","section":null},{"comment":"§2.A: the sentence defining Q_v for archimedean places reads “or R if v is non-archimedean”; the second occurrence should be “archimedean”.","section":null},{"comment":"§3.A, Remark 3.1: the discussion of the erroneous definition in André is helpful but could be shortened; the corrected characterisation via Lemma 3.2 is the one actually used later.","section":null},{"comment":"§4.C: the two corrections to André’s inequality are carefully justified; a short parenthetical note that the original asymptotic statements remain valid would reassure readers who only need the non-explicit theory.","section":null},{"comment":"§6.C: the numerical comparison with Habegger is clear, but the optimised constant 2^{27 n^{17}} for lines is stated without a reference or short derivation; a one-sentence sketch or citation would help.","section":null},{"comment":"Throughout: the notation σ(Ξ) for the height of a finite set is standard, yet a brief reminder that it coincides with the usual affine logarithmic height of the corresponding vector would aid non-specialists.","section":null}],"recommendation":"accept","confidential_remarks":"The manuscript is carefully written, self-contained, and fills a genuine gap. The geometric restrictions are methodologically necessary and do not undermine the stated theorems. I see no reason to delay publication; the paper is ready for acceptance after routine copy-editing."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"Martin has finally written down the explicit constants that everyone has been waving their hands about for Bombieri–André, and then used them on the simplest possible unlikely-intersection problem in tori. That is the news.\n\nWhat is new: Corollaries 4.5, 5.4 and 5.5 give fully explicit c1 (and improved c2 under super-strong non-triviality) for the height of points carrying non-trivial global relations among G-functions. Theorem 4.2 is the refined engine that tracks multiple independent relations and the ambient monomial space. The torus application (Theorem 6.1 / Corollary 1.3) is the first place the method has been written out for algebraic groups rather than Shimura varieties; for lines through the identity he gets a polynomial-in-degree bound that beats Habegger’s enormous constant for Galois degrees up to roughly 10^4 (or far higher if you take Habegger’s general-curve constant). The comparison in §6.C is honest.\n\nHe does the bookkeeping carefully: corrects the two small errors in André’s VII.3.5 inequality (s = max{1,#Sin} and the good-apparent-singularity hypothesis), introduces (rv)-global relations so the radii can be smaller than the radii of convergence without an extra ad-hoc step, and computes every input (radii, sizes, singularities of the log system) explicitly for the series log(1+aiX). Ax–Schanuel gives the functional independence, Hilbert-function bounds are standard, and the construction of the global relation of degree ≤ max{\\tau(K(s)),1} is clean once you accept the geometric restrictions.\n\nSoft spots are exactly the ones he flags: the bound depends on [K(s):Q], only Σ2 is treated, the curve must pass through a torsion point, and the ambient variety must be a line. Those are essential to getting Q-coefficients and controlled degree; they are not hidden. The paper does not claim more. No free parameters, no circularity.\n\nThis is for people who actually want to run the G-function machine or who need concrete numbers for small-degree points on lines. It deserves a serious referee; the math is checkable line-by-line. I would cite the explicit corollaries and the torus illustration.","headline":"Solid explicit constants for Bombieri–André plus the first clean written G-function application to tori; restrictions are real but openly stated and do not break the claims.","tokens_in":30502,"tokens_out":588,"would_cite":true,"duration_ms":8136,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["11G30","11G50","11J91"],"pacs":[],"model":"grok-4.5","headline":"Explicit height bounds for unexpected G-function relations give concrete height control on unlikely intersections of lines through the identity with rank-1 subgroups of tori.","keywords":["G-functions","height bounds","unlikely intersections","tori","Bombieri–André principle","global relations","explicit constants"],"falsifier":"Compute the actual Weil height of a concrete point of a line 1+a_i ξ in G_m^3 that lies in a rank-1 subgroup and has Galois degree 4 or 5; if that height exceeds the numerical constant 2708(H+1) of Corollary 1.3, the claimed bound is false.","tokens_in":30606,"feed_emoji":"📐","tokens_out":832,"duration_ms":8402,"temperature":0.7,"pith_summary":"The paper turns Bombieri–André’s asymptotic height bound for points with non-trivial global polynomial relations among G-function values into fully explicit inequalities, with constants written in terms of the number of functions, their sizes, radii of convergence, and the singularities of their differential equation. It then specialises the machinery to the logarithms of linear forms 1+a_i X, which are G-functions, and thereby obtains an explicit polynomial height bound for points of a line C through the identity in G_m^n that lie in the union of algebraic subgroups of codimension at least two. The bound depends on the number of complex places of the field of definition of the point and on the height of the coefficients of the line; for small Galois degree it improves the previously known explicit bound of Habegger, while for large degree it is weaker because the degree of the constructed global relation grows with the number of complex embeddings. The work also supplies the first fully written application of the G-functions method to unlikely intersections inside algebraic groups rather than Shimura varieties.","feed_headline":"Explicit heights for G-function relations bound torus intersections","feed_subtitle":"Polynomial control on points of lines through the identity that meet rank-1 subgroups, better than Habegger for small degree","key_machinery":"The corrected explicit inequality of André (Proposition 4.6) that converts a collection of linearly independent local relations of degree 1 into a global height lower bound; combined with the construction of an (r_v)-global relation of degree at most max{τ(K(s)),1} obtained by multiplying the linear forms attached to independent multiplicative characters at the complex places (Proposition 6.8).","core_discovery":"All points at which a super-strongly non-trivial (or merely non-trivial) (r_v)-global polynomial relation of degree δ holds among a fixed tuple of G-functions satisfy an explicit height bound of the shape 8 s λ^3/κ^2 (log(δ)+O(1)) (σ_y+ρ) (or the refined super-strong form involving Hilbert-function ratios and binomial constants c_6(η), c_7(η)). Specialised to the n+1 G-functions {1, log(1+a_i X)} attached to a line through the identity in G_m^n, this yields an explicit height bound for every point of the line that lies in Σ_2, polynomial of degree n in the number of complex places of its field of definition.","pith_inferences":[],"forward_implications":[],"fun_headline_variants":["Explicit height bounds for G-function relations on torus lines","Height bounds for unlikely intersections of lines with rank-1 tori","Bombieri-André constants made explicit for G-function polynomial relations","Explicit polynomial height bounds for lines meeting Σ₂ in G_m^n","G-function height estimates improve Habegger bounds for torus intersections"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"The curve must pass through the identity (or a torsion point) so that the local relations at complex places can be taken with rational coefficients and still work at all non-archimedean and real places; without that the degree of the global relation cannot be controlled solely by the number of complex embeddings.","fun_headline_variants_meta":{"raw":{"variants":["Explicit height bounds for G-function relations on torus lines","Height bounds for unlikely intersections of lines with rank-1 tori","Bombieri-André constants made explicit for G-function polynomial relations","Explicit polynomial height bounds for lines meeting Σ₂ in G_m^n","G-function height estimates improve Habegger bounds for torus intersections"]},"model":"grok-4.5","effort":"low","cost_usd":0.004446,"raw_usage":{"total_tokens":1281,"prompt_tokens":714,"num_sources_used":0,"completion_tokens":93,"cost_in_usd_ticks":44460000,"prompt_tokens_details":{"text_tokens":714,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":474,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":714,"tokens_out":93,"duration_ms":4937,"temperature":1.0,"reasoning_tokens":474,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-13T02:05:42.912092+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Compute the actual Weil height of a concrete point of a line 1+a_i ξ in G_m^3 that lies in a rank-1 subgroup and has Galois degree 4 or 5; if that height exceeds the numerical constant 2708(H+1) of Corollary 1.3, the claimed bound is false.","supporting_citations":[],"review_version":1}