REVIEW 3 major objections 4 minor 1 cited by
On the torsion values for sections of an elliptic scheme
T0 review · 3 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read An algebraic section's canonical height equals the area its Betti coordinates sweep out on the base curve.
desk verdict A substantial paper: it proves a clean height–Betti-measure integral for elliptic sections and adds effective Diophantine results, but the main counting argument leans on an unpublished definability/boundedness result. 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 Betti map sends a point on a fiber to the two real coefficients expressing its abelian logarithm as a combination of the fiber's periods. The load-bearing objects are the closed two-form $d\beta_1\wedge d\beta_2$ on the elliptic surface minus the singular fibers, pulled back to the base by the section, and the second-order differential operator $\Xi$ whose vanishing characterizes torsion sections and which controls intersection multiplicities. The counting step uses the fact that the Betti coordinates of an $n$-torsion point are rational with denominator dividing $n$, together with a tame definability and boundedness result for the Betti map and a lattice-point counting theorem for definable sets: these convert the count of torsion points into the Lebesgue area of the Betti-coordinate image, proving the integral formula.
What would settle it
Integrate the right-hand side of (3.6) for an explicit section with independently known height, such as $\sigma(\lambda)=(2,\sqrt{2(2-\lambda)})$ on the base $t^2+2$: the formula requires the value to be exactly $1/2$, and any numerical discrepancy after handling branches and singular terms would disprove Theorem 3.2. More generally, any algebraic section whose intersection-theoretic canonical height disagrees with the computed Betti-area integral would falsify the main identity.
Extended reading notes
Core claim
For a non-torsion algebraic section $\sigma$ of an elliptic scheme over a curve, the paper establishes that the canonical height $\hat h(\sigma)$ equals the integral of the pulled-back two-form $\sigma^*(d\beta_1\wedge d\beta_2)$, computed over the base with the bad fibers removed. The proof counts $n$-torsion values of $\sigma$ twice: intersection products with the zero section give a limit $\hat h(\sigma)$, while the Betti-coordinate description of torsion gives a limit equal to the Betti area; the two limits must coincide. Along the way the paper shows the torsion values are equidistributed with respect to that Betti measure, that the typical multiplicity of a torsion value is $2$ with only finitely many higher-multiplicity exceptions, and that the Betti measure is exactly the closed current appearing in dynamical equidistribution theorems. In the final part, the multiplicity problem is connected to Diophantine approximation over function fields and yields an effective, valuation-independent bound $|\xi|_v\le C\, H(\xi)^{\varepsilon}$ for abscissae of points on elliptic curves.
Load-bearing premise
The counting argument assumes that the Betti map, restricted to the base minus the bad fibers, is tamely definable and bounded, so that the number of rational coordinate pairs with denominator $n$ in its image follows the Lebesgue-area law; this is quoted from an unpublished preprint, and a failure near $0,1,\infty$ would break the equidistribution step and the integral formula.
Editorial extensions
If this is right
- The canonical height of any algebraic section is a rational number, so it can in principle be evaluated by numerical integration of the Betti form once the periods of the family are known.
- For a non-torsion section, the torsion values are distributed over the base by the Betti measure, and the number of points where $\sigma$ has order dividing $n$ grows like $\hat h(\sigma)n^2$ with the expected multiplicity $2$.
- The exceptional set where a torsion value is attained with multiplicity greater than expected is finite; over a finitely generated group of sections the union of all such exceptional sets is finite and uniformly bounded.
- The Betti measure coincides with the dynamical equidistribution current, so torsion-order sequences and general height-zero sequences of points share the same limiting distribution on the base.
- For elliptic curves over function fields, an effective analogue of the classical theorem on integral points holds: abscissae satisfy $|\xi|_v \le C\,H(\xi)^{\varepsilon}$ with $C$ independent of the valuation $v$.
Reading between the lines
- Beyond the paper, one could test the formula as a computational tool in the Legendre family by computing both sides for sections $\sigma_k=[k]\sigma_0$: the theorem predicts the Betti integral scales quadratically in $k$, which could be checked numerically without knowing a Mordell-Weil basis.
- Beyond the paper, the remark that the naive higher-dimensional analogue fails because the $[n]$-pullback scales the form by $n^{2g}$ suggests that any extension to abelian schemes of dimension $g>1$ would need a different normalization or a different measure, and that multiplicities there may not be uniformly bounded.
- Beyond the paper, the valuation-independent effective bound could be used to make the finiteness theorem for high-multiplicity torsion points quantitative over a fixed finitely generated section group, by replacing the tame definability input with explicit counting where the field and group are explicit.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript studies sections of an elliptic scheme over an affine base curve and analyzes the locus where the section takes torsion values. The central object is the Betti map of a section; the paper proves an integral formula (Theorem 3.2) expressing the canonical height of a non-torsion section as the integral of the pulled-back form dβ1 ∧ dβ2 over the base, and shows that torsion points of order dividing n are asymptotically distributed according to this measure. The proof has two independent counting ingredients: an algebraic intersection count showing that the number of n-torsion points is asymptotic to n² times the canonical height, and an o-minimal counting argument, using a result of Barroero–Widmer, showing the same count is asymptotic to the integral of the Betti form. The paper also proves finiteness results for the points where the torsion multiplicity is larger than expected, compares the Betti measure with the DeMarco–Mavraki dynamical current, and derives an effective Siegel-type bound for the v-adic size of points on elliptic curves over function fields.
Significance. If the quoted definability and boundedness results are available, Theorem 3.2 is a valuable bridge between arithmetic height and the analytic Betti map, and the distribution statement for torsion points is a genuine strengthening of the qualitative density results known previously. The comparison in Section 4 with the DeMarco–Mavraki current is non-circular: it proves equality of the currents via a uniqueness property in Proposition 4.3 rather than assuming the height formula, and the alternative proof in Section 4.1 shows how Theorem 3.2 follows from that comparison combined with Wirtinger’s formula. The multiplicity finiteness results and the effective quasi-integral-point theorem in Section 5 are also substantial. The main caution is the reliance of the counting proof on the unpublished preprint [10] for definability and, crucially, boundedness of the Betti map; this is the single point on which the central integral formula currently rests.
major comments (3)
- [§3.2, proof of Theorem 3.2(b)] The proof applies Lemma 3.5 (Barroero–Widmer) to sets built from graphs of branches of the Betti map. This requires, in addition to definability, that each branch Bj be bounded on its piece Yj. The manuscript asserts boundedness by quoting [10, Proposition 4], but [10] is an unpublished preprint and no argument is given in the present text. If a branch is unbounded near a point of r^{-1}(S), the sets A_j^m obtained from the Hardt trivialization may be unbounded, the counts α_{n,j}^m need not be comparable to n² μ(A_j^m), and equations (3.13)–(3.14) lose their justification. The estimate (3.18) in Remark 3.7 bounds the pulled-back 2-form dβ_1^t ∧ dβ_2^t, not the Betti coordinates themselves, so it does not repair the missing boundedness. I ask the authors to prove the needed boundedness statement, or to cite a precise and verified published statement, or to restructure the counting argument so that it does not require boundedness of the Betti map.
- [§3.2, proof of Theorem 3.2(a)] The proof that δ_n + s_n = O(1) is only sketched at one point: for a bad-reduction fiber one reduces to the case m = 1 by replacing σ by mσ, but the text does not explicitly write the relation between the local intersection numbers of nσ and the corresponding multiples of mσ for the original section. The formal-group argument leading to (3.10) is sound for the normalized section meeting the special fiber at the origin, but the reduction justifying the same conclusion for all n with m | n should be stated. Since this boundedness is what identifies the limit of A_n/n² with the intersection-theoretic canonical height, the step should be made fully explicit.
- [§2, Theorem 2.8 and Proposition 2.4] The proof of Theorem 2.8 asserts that the function Ξ(σ̃) has no essential singularities on a complete model of B by 'easy growth estimates', and then concludes it is rational. This assertion is load-bearing for the multiplicity bound and for the subsequent finiteness results, but no growth estimate is actually displayed. Similarly, Proposition 2.4 uses definability of the Betti map quoted from the unpublished [10] near the boundary of B; this is acceptable only if the precise statement in [10] is supplied. The authors should either prove the growth assertion or give a complete reference to a published argument.
minor comments (4)
- [§2, Definition 2.1] The phrase 'It may be continued to all of B(C), with monodromy transformation that we forget about' is potentially confusing, since the Betti map is not globally single-valued; please clarify that the continuation is multivalued and that the differentials dβ1, dβ2 are the globally well-defined objects.
- [§3, Example 3.4] The computation for the section σ(λ) = (2, √(2(2−λ))) would benefit from a precise description of the branch cuts and the 'segment [2,∞]' topology on B; without this, the displayed equality relating 1/4, the height, and the integral is hard to verify.
- [§3, equation (3.7)] The notation β_i is overloaded: in (3.7) it denotes the Betti coordinates of the section σ, while elsewhere β denotes the Betti map on the total space. It would help to write β_i∘σ explicitly in the integrand.
- [Throughout] There are several typographical slips, e.g. 'Morevoer' in the Introduction and 'hanece' in Remark 2.11; these should be corrected in revision.
Circularity Check
No significant circularity: the height integral and the Betti-measure counting are derived from independent inputs, and the cited definability results are external black boxes.
full rationale
The paper's central theorem (Theorem 3.2) compares two independently defined quantities: the canonical height hhat(sigma), defined via intersection theory, and the integral of sigma^*(d beta_1 wedge d beta_2), defined from the periods and the Betti map. Neither quantity is defined in terms of the other, and no parameter is fitted to the target. The counting quantity A_n is introduced independently as the number of base points where sigma(t) is n-torsion, and the proof shows two separate limits, one equal to the height and one equal to the Betti-measure integral. The use of definability and boundedness of the Betti map is explicitly attributed to the external preprint [10] by Jones and Schmidt; this is a genuine external result, not a self-citation, and any failure of that quoted theorem would be a correctness risk, not a circularity. The finiteness of fibers of the Betti map is quoted from [7], a published paper with overlapping authors, but it is an independent geometric statement about Betti fibers, not an assertion of the height formula or of the measure equality. The comparison with DeMarco-Mavraki's current in Section 4 is supported by a self-contained uniqueness argument (Proposition 4.3) and an explicit calculation (Remark 4.2), so it does not reduce to a renaming or to a citation. Overall, the derivation chain is not circular by construction; it relies on external theorems as black boxes, which is standard mathematical practice.
Assumptions & free parameters
assumptions (7)
- domain assumption The Betti map of the Legendre scheme is piecewise definable and bounded in the o-minimal structure R_an,exp (Jones-Schmidt [10]).
- domain assumption Manin's theorem: a section whose Betti map is constant is torsion.
- domain assumption Andre's theorem on linear independence of Betti maps of independent sections.
- domain assumption Silverman-Tate bounded height theorem for torsion values of a non-torsion section.
- domain assumption DeMarco-Mavraki equidistribution theorem for small points (Theorem 4.1).
- domain assumption Barroero-Widmer lattice point counting (Lemma 3.5).
- standard math Standard intersection theory and formal group results for elliptic surfaces.
Cite this review
Pith. "Pith review of On the torsion values for sections of an elliptic scheme." pith.science (2026). https://pith.science/paper/F2NFEUVK
@misc{pith2026190901253,
author = {Pith},
title = {Pith review of: On the torsion values for sections of an elliptic scheme},
year = {2026},
howpublished = {\url{https://pith.science/paper/F2NFEUVK}},
note = {Machine review of arXiv:1909.01253}
}
abstract
We shall consider sections of an elliptic scheme $\mathcal{E}$ over a(n affine) base curve $B$, and study the points of $B$ where the section takes a torsion value. In particular, we shall relate the distribution in $B$ of these points with the canonical height of the section, proving an integral formula involving a measure on $B$ coming from the so-called Betti map of the section. We shall show that this measure is the same appearing in dynamical issues related to the section. This analysis will also involve the multiplicity with which a torsion value is attained, which is an independent problem. We shall prove finiteness theorems for the points where the multiplicity is higher than expected. Such multiplicity has also a relation with Diophantine Approximation and quasi-integral points on $\mathcal{E}$ (over the affine ring of $B$), and in the last part of the paper we shall exploit this viewpoint, proving an effective result in the spirit of Siegel's theorem on integral points.
Forward citations
Cited by 1 Pith paper
-
Transversality of sections on elliptic surfaces with applications to elliptic divisibility sequences and geography of surfaces
On elliptic surfaces, all but finitely many multiples of a non-torsion section meet the zero section transversely, yielding new transversality results and surfaces with fixed geometric genus and unbounded K^2.
Reference graph
Works this paper leans on
-
[10]
Pfaffian definitions of Weierstrass elliptic functions
G. Jones and H. Schmidt, Pfaffian definitions of Weierstrass ellipt ic func- tions, preprint, https://arxiv.org/abs/1709.05224v3
-
[1]
Y. Andr´ e, P. Corvaja, U. Zannier, On the Betti map associate d to a section of an abelian scheme, https://arxiv.org/abs/1802.03204
-
[2]
E. Bombieri and A.J. van der Poorten, Some quantitative results related to Roth’s Theorem, J. Australian Math. Soc. 45 (1988), 233–248
work page 1988
-
[3]
F. Barroero, M. Widmer, Counting lattice points and o-minimal str uctures. Int. Math. Res. Not. IMRN (2014), no. 18, 4932-4957
work page 2014
-
[4]
A. Beauville, Complex Algebraic Surfaces , London Mathematical Society student texts, Cambridge University Press, 1996
work page 1996
-
[5]
Bierstone, Milman, Semianalytic and subanalytic sets, Publ
E. Bierstone, Milman, Semianalytic and subanalytic sets, Publ. Mat h. IHES 47, 5-42 (1988)
work page 1988
-
[6]
Van den Dries, Tame Topology and O-minimal Structures, 150 184
L. Van den Dries, Tame Topology and O-minimal Structures, 150 184. Cam- bridge University Press, 1998
work page 1998
-
[7]
P. Corvaja, D. Masser, U. Zannier , Torsion Hypersurfaces on Abelian Schemes and Betti Coordinates , Math. Annalen 371 (2018), 1013-1045. 43
work page 2018
Show all 27 references
-
[8]
Corvaja and U
P. Corvaja and U. Zannier, Poncelet games, Manin’s kernel theo rem, the Betti map and torsion in group schemes, preprint
-
[9]
De Marco, N
L. De Marco, N. Mavraki, Variation of canonical height and equidis tribu- tion, to appear on American Journal of Math
-
[11]
Lawrence, U
B. Lawrence, U. Zannier, On the p-adic distribution of torsion values for a section af an abelian scheme, https://arxiv.org/abs/1908.09050
1908 arXiv
-
[12]
Manin, Rational Points of Algebraic Curves over Function Fie lds, Izv
Ju. Manin, Rational Points of Algebraic Curves over Function Fie lds, Izv. Akad. Nauk SSSR Ser. Mat. 27:6 (1963), 1395-1440
1963
-
[13]
Mason, Diophantine equations over function fields, London Math
R.C. Mason, Diophantine equations over function fields, London Math. Soc. Lecture Note Series 96, Cambridge 1984
1984
-
[14]
diophantine approximatio n
C.F. Osgood, Effective bounds on the “diophantine approximatio n” of al- gebraic functions over fields of arbitrary characteristic and applic ations to differential equations, Indag. Math. 37 (1975), 105–119
1975
-
[15]
Osgood, Sometimes effective Thue-Siegel-Roth-Schmidt-N evanlinna bounds, or better, J
C.F. Osgood, Sometimes effective Thue-Siegel-Roth-Schmidt-N evanlinna bounds, or better, J. Number Th. 21 (1985), 347–389
1985
-
[16]
Schmidt, On Osgood’s effective Thue Theorem for algebraic f unc- tions, Comm
W.M. Schmidt, On Osgood’s effective Thue Theorem for algebraic f unc- tions, Comm. Pure and Applied Math. 29 (1976), 759–773
1976
-
[17]
Serre, Lectures on the Mordell-Weil Theorem, Aspects of Mathematics, Vieweg+Teubner Verlag, 2013
J.-P. Serre, Lectures on the Mordell-Weil Theorem, Aspects of Mathematics, Vieweg+Teubner Verlag, 2013
2013
-
[18]
Silverman, The Arithmetic of Elliptic Curves , Graduate Text in Math- ematics, Springer Verlag 1986
J.H. Silverman, The Arithmetic of Elliptic Curves , Graduate Text in Math- ematics, Springer Verlag 1986
1986
-
[19]
Silverman, Advanced topics in the arithmetic of elliptic curves , Grad- uate texts in mathematics, Springer Verlag, 1994
J.H. Silverman, Advanced topics in the arithmetic of elliptic curves , Grad- uate texts in mathematics, Springer Verlag, 1994
1994
-
[20]
Shioda and M
T. Shioda and M. Schuett, Elliptic Surfaces, ArXiv e-prints , July 2009
2009
-
[21]
Tate, Variation of the canonical height of a point depending o n a param- eter, American Journal of Mathematics , 105(1):287–294, 1983
J. Tate, Variation of the canonical height of a point depending o n a param- eter, American Journal of Mathematics , 105(1):287–294, 1983. 44
1983
-
[22]
Ulmer and G
D. Ulmer and G. Urzua, Tansversality of sections on elliptic surfa ces with applications to elliptic divisibility sequences and geography of surface s, https://arxiv.org/abs/1908.02208
1908 arXiv
-
[23]
Voloch, Siegel’s Theorem for complex function fields, Proc
J.F. Voloch, Siegel’s Theorem for complex function fields, Proc. A mer. Math. Soc. 121 (1984), 1307–1308
1984
-
[24]
Wang, An effective Roth’s theorem for function fields , Rocky Mountain J
J. Wang, An effective Roth’s theorem for function fields , Rocky Mountain J. Math. 26, 3 (1996), 1225-1234
1996
-
[25]
Zannier, Some Problems of Unlikely Intersections in Arithmetic and Geometry, Annals of Math
U. Zannier, Some Problems of Unlikely Intersections in Arithmetic and Geometry, Annals of Math. Studies 181, Princeton U. Press 2012
2012
-
[26]
Zannier, Lecture Notes on Diophantine Analysis, with an appendix by F
U. Zannier, Lecture Notes on Diophantine Analysis, with an appendix by F. Amoroso, Publications of the Scuola Normale Superiore, Scuola Normale Superiore, 2015
2015
-
[27]
Zimmer, Die N´ eron-Tate’schen quadratischen Formen auf der ratio- nalen Punktgruppe einer elliptischen Kurve, J
H.G. Zimmer, Die N´ eron-Tate’schen quadratischen Formen auf der ratio- nalen Punktgruppe einer elliptischen Kurve, J. Number Theory 2 (19 70), 459–499. 45
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.