REVIEW 3 major objections 5 minor 10 references
Algebraic conditions for the positivity of sectional curvature
T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read This paper establishes that, for 4-dimensional curvature operators, sectional positivity is equivalent, on a dense open set, to all real roots of a single discriminant polynomial being positive.
desk verdict A plausible discriminant criterion for sectional positivity in dimension 4, with a clean forward direction but a converse that rests on an unproven genericity claim. 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 object carrying the argument is the two-variable characteristic polynomial $p(x,y)=\det(R-xI-yK)$, where $K$ is the 4-volume form acting on 2-forms, and the single-variable polynomial $q(x)=\operatorname{disc}_y(p(x,y))$ obtained by taking the discriminant in $y$. A pair $(x_1,y_1)$ is a critical point of the Lagrangian $vRv-x(vIv-1)-y(vKv)$ exactly when $y_1$ is a multiple root of the $y$-polynomial $\det(R-x_1I-yK)$; the discriminant $q$ records precisely those $x_1$ values, and the supplementary discriminant $\operatorname{disc}_x(q)$ is used to exclude false real roots coming from non-real multiple roots. The work it does is to convert an extremal problem over a compact Grassmannian into a finite algebraic sign test on the real roots of a single polynomial.
What would settle it
Search for a 4-dimensional Riemann tensor $R$ with $\operatorname{disc}_x(q)\neq 0$ for which some real root $x_1$ of $q$ has all multiple roots of $\det(R-x_1I-yK)$ non-real; finding one would break the equivalence in Theorem 1.1.
Extended reading notes
Core claim
The central assertion is Theorem 1.1: for $n=4$, let $p(x,y)=\det(R-xI-yK)$ and $q(x)=\operatorname{disc}_y(p)$. The set of real roots of $q$ contains the set of critical values of the sectional curvature of $R$; if $\operatorname{disc}_x(q(x))$ is nonzero, the sets are equal. Consequently all real roots of $q$ positive (nonnegative) implies $R$ is sectionally positive (nonnegative), and under the nonvanishing discriminant condition the implication is an equivalence. The proof passes through a Lagrange-multiplier description of critical points of $vRv$ on the Grassmannian of oriented 2-planes, then eliminates the vector $v$ by studying when the first nonzero homogeneous part of $\det(R-xI-yK)$ changes sign behavior. The paper also shows the analogous higher-dimensional discriminant construction collapses: for $n\ge 5$ the resulting $q$ is identically zero.
Load-bearing premise
The converse of the main theorem rests on the claim that, whenever $\operatorname{disc}_x(q)$ is nonzero, every real root $x_1$ of $q$ arises from a real multiple root of $\det(R-x_1I-yK)$; the paper's proof of this lemma appeals to a genericity assertion without deriving it algebraically.
Editorial extensions
If this is right
- Sectional positivity in dimension 4 can be certified by checking the signs of the real roots of $q$, using exact real-root algorithms, without searching over 2-planes.
- When $\operatorname{disc}_x(q)\neq 0$, the roots of $q$ give the exact maximum and minimum of sectional curvature, so curvature bounds are computable from $q$ alone.
- The main theorem recovers the known characterization that in dimension 4 every sectionally positive operator is strongly positive, meaning $R-yK$ is positive definite for some $y$.
- The discriminant method does not extend naively to $n\ge 5$: the analogous $q$ is identically zero, so a different algebraic object would be needed.
- The criterion gives a purely algebraic route to checking a strong geometric condition, with no need to solve the constrained critical-point equations geometrically.
Reading between the lines
- Because $q$ has bounded degree in $x$ for $n=4$, the sign test is finite and likely implementable in exact arithmetic; one could stratify the space of 4-curvature operators by the vanishing of $\operatorname{disc}_x(q)$ and verify the theorem computationally on each stratum.
- The same discriminant-of-a-pencil idea could apply to other varieties defined by quadratic constraints, replacing $K$ with the defining quadratic forms of a homogeneous space; critical values of a quadratic form restricted to such a variety would then have an algebraic certificate.
- The failure in $n\ge 5$ suggests that multi-parameter discriminants or resultants of the family $\det(R-xI-\sum y_p K_p)$ may still produce useful algebraic certificates, but they cannot be built from the single multivariate discriminant alone.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes an algebraic criterion for sectional positivity of curvature operators in dimension 4. For a curvature operator R on Λ^2, with I the identity and K the Hodge star operator, it defines p(x,y)=det(R−xI−yK) and q(x)=disc_y(p). The main theorem (Theorem 1.1) claims that every critical value of the sectional curvature function on the Grassmannian of oriented 2-planes is a real root of q, and that when disc_x(q)≠0 the two sets coincide. Consequently, sign conditions on the real roots of q give sufficient, and in the generic case necessary, conditions for sectional positivity. The paper also sketches a higher-dimensional analogue and its failure (Section 4), and applies the method to recover Thorpe's trick in dimension 4 (Section 5).
Significance. If the characterization is correct, it would provide a finite algebraic sign test for sectional positivity in dimension 4, replacing a two-quantifier description by a univariate discriminant. The construction is self-contained, and the forward direction (Theorem 3.7) is clean and convincing. The paper also identifies a genuine obstruction to a naive discriminant approach in higher dimensions. However, the converse half of the main theorem is not rigorously established: it depends on Lemma 3.10, whose proof rests on unsupported genericity and stability assertions, and on Lemma 3.9, whose Sylvester-matrix argument is not verifiable as displayed. The paper's principal claim is therefore currently conditional.
major comments (3)
- [Section 3, Lemma 3.10] Lemma 3.10 is the load-bearing step for the converse of Theorem 1.1, but its proof is incomplete. The proof asserts without derivation that the property 'for a real simple root x1 of q, all multiple roots of f(y)=det(R−x1I−yK) are non-real' is open in the operator R, and that it is non-generic ('Generically, this does not happen'). The openness step is not justified: although the implicit function theorem gives a nearby simple root x1(R') of q_R', it is not shown that the multiple roots of f_R'(y)=det(R'−x1(R')I−yK) remain non-real and that no real multiple root is created. The genericity step is only a dimension heuristic; it requires a proof that the resultant of disc_y(p) and the first subdiscriminant is not identically zero and that the bad set does not intersect {disc_x(q)≠0}. Without Lemma 3.10, Theorem 3.11 and the if-and-only-if in Theorem 1.1 are not established.
- [Section 3, Lemma 3.9] Lemma 3.9 is not verifiable as written. The displayed submatrix S2 of the Sylvester matrix is garbled, and the argument that every 2-minor M_ab of S2 is divisible by x^2 is not actually carried out in the text. Since Lemma 3.9 is used in Theorem 3.11 to conclude a10≠0, this is a load-bearing gap. A clean statement of the Sylvester matrix and a complete proof of the divisibility claim are needed.
- [Section 5, Theorem 5.2] In the proof of Theorem 5.2, the step 'Due to disc_x(q(x))≠0, the roots of the polynomial are always distinct (see Lemma 3.9), regardless of y' is a non sequitur. Lemma 3.9 only rules out a10=0 at pairs (x,y) where y is a multiple root of the y-polynomial; it does not imply that for every fixed y the polynomial det(R−xI−yK) has no multiple roots in x. The proof of Theorem 5.2 is therefore incomplete as written. This does not affect the main theorem directly, but it indicates that the stated applications rely on claims beyond those actually proven.
minor comments (5)
- [Section 3, Lemma 3.2, proof] The identity P_k(x,1)=a_{k0}T_{K'}(x) used in the proof of Lemma 3.2 is not correct in general; the correct relation is P_k(x,1)=a_{k0}T_{K'}(−x) (up to sign conventions). The criterion in the lemma is invariant under this change, so the statement survives, but the proof should be corrected.
- [Abstract and Section 1] The paper advertises a 'complete characterization for a dense open subset' but does not prove that {disc_x(q)≠0} is nonempty. Since this set is the complement of an algebraic hypersurface, density would follow from a single example; such an example should be supplied.
- [Throughout] There are many typographical and formatting issues, including 'dimentional', 'cur vature', inconsistent spacing around operators, and in particular the displayed Sylvester matrix in Lemma 3.9, which is unreadable and must be typeset properly.
- [Section 4, Theorem 4.2] The phrase 'non-zero measure' is informal and should be clarified. The proof that q vanishes on a non-null set relies on dimension estimates that are not given; a precise statement about the dimension of the bad set is needed.
- [Section 2, Theorem 2.1] The claim that a polynomial with alternating coefficients is monotone and of constant sign on (−∞,0] is stated without proof. A short argument using the sign of each term would make the proof self-contained.
Circularity Check
No significant circularity: the main construction is self-contained and does not feed the target result back into its inputs.
full rationale
The paper's central object, q(x)=disc_y(det(R-xI-yK)), is constructed directly from the Riemann operator R, the identity I, and the volume form K, and the critical values of sectional curvature are characterized via the Lagrangian LR(v,x,y)=vRv-x(vIv-1)-y(vKv). The inclusion Theorem 3.7 is proved from the Lagrange multiplier conditions, not assumed, and the converse Theorem 3.11 invokes Lemma 3.9 and Lemma 3.10 as auxiliary algebraic statements rather than as fitted or predefined equivalences. No parameter appearing in the final criterion is fitted to the sectional curvature values, and no load-bearing result is imported solely from the author's own prior work; the citations to Singer-Thorpe and Puettmann supply the standard Lagrangian setup, but the algebraic construction is carried out in the paper. The proof of Lemma 3.10 contains a genericity assertion ('Generically, this does not happen') that is not fully justified, and Lemma 3.9's Sylvester-matrix display is garbled; these are correctness or completeness concerns, not circularity. There is no exhibited step in which the claimed prediction reduces by construction to its own input.
Assumptions & free parameters
assumptions (4)
- standard math Standard properties of discriminants and Sylvester matrices, including behavior when the top coefficient is constant.
- standard math The Lagrangian critical point method on the compact Grassmannian of oriented 2-planes characterizes sectional curvature bounds.
- ad hoc to paper For operators with disc(q) nonzero, non-real multiple roots of f(y)=det(R-x1I-yK) form a measure-zero set and can be excluded.
- domain assumption The curvature operator R is a real symmetric operator on the space of 2-forms, and K is the Hodge volume form with three positive and three negative eigenvalues.
Cite this review
Pith. "Pith review of Algebraic conditions for the positivity of sectional curvature." pith.science (2026). https://pith.science/paper/QNL7PKPT
@misc{pith2026190806476,
author = {Pith},
title = {Pith review of: Algebraic conditions for the positivity of sectional curvature},
year = {2026},
howpublished = {\url{https://pith.science/paper/QNL7PKPT}},
note = {Machine review of arXiv:1908.06476}
}
abstract
We examine algebraic conditions for the sectional positivity of the Riemann curvature operator. We describe sufficient conditions for dimension $n=4$, and complete characterization for a dense open subset of the space of operators in dimension $4$. We also briefly examine higher-dimentional curvature operators.
Reference graph
Works this paper leans on
-
[1]
R. Bettiol, M. Kummer, and R. Mendes. Convex algebraic geometry of curvature operators (to appear)
-
[2]
Marvin Marcus, Determinants of Sums , College Mathematics Journal, 2(21), 1990, 130–135
work page 1990
-
[3]
Saugata Basu, Richard Pollack and Marie-Fran¸ coise Roy, Algorithms in Real Algebraic Geometry, second edition, Algorithms and Computations in Mathematics, Volume 10, Springer, 2016
work page 2016
-
[4]
Ruslan Sharipov, Multiple discriminants and critical values of a multivaria te polyno- mial, arXiv:1508.00551
-
[5]
Renato G. Bettiol, Ricardo A. E. Mendes Strongly positive curvature, Annals of Global Analysis and Geometry, 53(3), 2018, 287–309
work page 2018
-
[6]
Thorpe The zeroes of non-negative curvature operators, J
John A. Thorpe The zeroes of non-negative curvature operators, J. Differential Geom., 5(1-2), 1971, 113–125
work page 1971
-
[7]
Thorpe Erratum:”The zeroes of non-negative curvature operators” , J
John A. Thorpe Erratum:”The zeroes of non-negative curvature operators” , J. Differ- ential Geom., 5, 1971, 113–125. J. Differential Geom., 11, 1976, 315
work page 1971
-
[8]
I.M Singer and J. A. Thorpe. The curvature of 4-dimensional Einstein spaces . In Global Analysis (Papers in Honor of K. Kodaira) , 355-365. Univ. Tokyo Press, Tokyo, 1969
work page 1969
Show all 10 references
-
[9]
Invent Math., 138 (1999) 631-684
Thomas P¨ uttmann Optimal pinching constants of odd-dimensional homogenous spaces. Invent Math., 138 (1999) 631-684
1999
-
[10]
1882 E-mail address : dan.fodor52@yahoo.com F aculty of Mathematics, Alexandru Ioan Cuza University, Ia s ¸i, Romania
Thomas Muir, A treatise on the theory of determinants , Macmillan & Co. 1882 E-mail address : dan.fodor52@yahoo.com F aculty of Mathematics, Alexandru Ioan Cuza University, Ia s ¸i, Romania
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.