Serendipitous Syzygies of Scattering Amplitudes

David A. Kosower; Sebastian P\"ogel

REVIEW 3 major objections 5 minor 14 references

Serendipitous Syzygies of Scattering Amplitudes

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read For n≥5, one-loop all-plus Yang-Mills amplitudes obey (n−1)!/2−2 linear relations, leaving exactly two independent partial amplitudes.

desk verdict Solid paper with a nice structural theorem and useful explicit syzygies; the abstract overclaims 'only two' as proven when the exact count is verified only through n=7/8 and conjectured beyond. read the letter →

arxiv 2505.14857 v1 pith:WPD7L7BU submitted 2025-05-20 hep-th hep-ph

David A. Kosower , Sebastian Pögel This is my paper · ORCID

classification hep-thhep-ph MSC 13D0213P1081T18

keywords scatteringamplitudesYang-Millstheoryall-plusMHVcolor-orderedsyzygieslinearrelationsone-loop

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper establishes that the color-ordered one-loop all-plus amplitudes of Yang–Mills theory, and the tree-level maximally helicity-violating (MHV) amplitudes, are far more redundant than their $(n-1)!/2$ distinct permutations suggest: for $n\ge 5$ there are $(n-1)!/2 - 2$ independent linear relations with polynomial coefficients, leaving exactly two independent partial amplitudes. The reason is general: any set of rational spinor functions sharing the same phase under spinor rescalings can be multiplied by one common monomial, re-expressed in momentum invariants, and after clearing denominators lands in the two-dimensional space spanned by the invariants and the Levi-Civita pseudoscalar, so any triple of amplitudes automatically satisfies a polynomial identity. The paper constructs the relations explicitly through seven points for all-plus amplitudes and eight points for MHV amplitudes, gives generators from which all relations follow by permutation, and provides two systematic methods for finding them. The result matters because it sharply reduces the data needed to compute these amplitudes and exposes a hidden algebraic structure whose low degree the paper traces to factorization rather than to the general argument alone.

What carries the argument

The central object is a syzygy: a vector of polynomials $P_n[\sigma]$ in the Mandelstam invariants such that $\sum_\sigma P_n[\sigma] A_n(\sigma)=0$. The load-bearing identity is the reduction of each phase-adjusted rational amplitude to $P_a(s_{ij}) + P_b(s_{ij})\epsilon_b$, which places all amplitudes in a two-dimensional space and yields the triple-cross-product syzygies of Eq. (3.8). Factorization at complex collinear poles $\langle i j \rangle$ converts the syzygy condition into recursive equations involving lower-point amplitudes, and these are solved with Gröbner-basis and syzygy computations. A permutation operation then generates the full set of relations from a small number of seed vectors, often a single one.

What would settle it

For $n=8$ all-plus amplitudes, evaluate the phase-adjusted amplitude vector on $2520$ generic rational momentum configurations and compute the nullity of the resulting matrix; the claim predicts nullity $2518$, so a smaller nullity that cannot be restored by higher-degree syzygies would disprove the reduction to two independent amplitudes.

Watch

Extended reading notes

Core claim

The central claim is that for $n\ge 5$ the vector of $(n-1)!/2$ color-ordered one-loop all-plus partial amplitudes has $(n-1)!/2 - 2$ independent syzygies, so only two amplitudes remain independent, and the same counting holds for tree-level MHV amplitudes through $n=8$. The proof rests on a reduction: after multiplying every partial amplitude by one common spinor factor that removes its spinor phase weight, and clearing the denominators that appear when re-expressing spinor products in momentum invariants, a rational amplitude without branch cuts takes the form $P_a(s_{ij}) + P_b(s_{ij})\epsilon_b$, where $\epsilon_b$ is the Levi-Civita pseudoscalar of a fixed momentum basis. Any three such expressions admit a polynomial identity built from the pairwise cross-products of their $P_a,P_b$ coefficients, which implies $n-2$ independent relations among $n$ amplitudes. The paper verifies for $n=7$ all-plus and $n=8$ MHV amplitudes that the syzygies through degree two exhaust these relations, and conjectures the exhaustion holds for all $n$.

Load-bearing premise

The argument assumes the amplitudes are purely rational functions of spinors with a common spinor phase weight and no branch cuts, so after one shared rescaling they live in the two-dimensional space of momentum invariants plus the Levi-Civita pseudoscalar; generic loop amplitudes with cuts, or amplitudes in other spacetime dimensions, do not satisfy this.

Editorial extensions

If this is right

For $n\ge 5$, the one-loop all-plus partial amplitudes are spanned by two functions, with explicit syzygies supplied through $n=7$.
Tree-level MHV amplitudes through $n=8$ satisfy the same two-amplitude counting, with the standard color identities and the momentum-dependent BCJ identities appearing as the degree-zero and degree-one layers of one syzygy module.
The two methods, exact numerical linear algebra over finite fields and factorization-based syzygy computation, give a general search procedure for polynomial relations among rational amplitudes.
Because the relations are exact and of low degree, they can reduce the number of amplitudes a numerical computation must evaluate, and the paper conjectures that only relations of degree at most two are needed at every multiplicity.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The same counting should apply to any helicity configuration whose amplitudes are rational in spinors with a common phase weight, such as the one-loop single-minus family; the paper notes multiparticle poles make those relations harder to find, so a numerical rank test there would separate the general mechanism from the factorization-based shortcut.
The empirical fact that one seed vector generates all relations under permutation suggests the syzygy module is cyclic as a representation of the symmetric group; proving cyclicity for all $n$ would give an all-multiplicity description and may connect to the KLT-kernel redundancy mentioned in the conclusions.
For practical computation, one could precompute the two independent amplitudes and reconstruct any partial amplitude by polynomial combination, although the growth of the polynomial coefficients with $n$ will determine whether this is actually cheaper than evaluating all permutations.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The genuinely new thing here is the Sec. III theorem: any set of rational functions of spinors with a common phase weight, after clearing denominators and phase, lives in the two-dimensional space spanned by 1 and the Levi-Civita pseudoscalar. So any triplet has a polynomial identity, and m amplitudes admit m−2 relations. That is a clean, general statement and it is proven properly. The paper then backs it up with two independent computational approaches—finite-field linear algebra and factorization-plus-syzygies—which cross-validate each other. The explicit syzygy generators for all-plus n=5,6,7 and MHV n=6,7,8 are a useful concrete payoff, and the ancillary files appear complete. I agree with the reader's high-level verdict: this is a worthwhile paper and the main existence result is solid.

The soft spot is the gap between the abstract and the theorem. The abstract says 'there are (n−1)!/2−2 relations for n≥5, leaving only two independent color-ordered amplitudes.' The theorem in Sec. III proves at most two independent amplitudes, not exactly two. The exact count is verified numerically for n≤7 all-plus and n≤8 MHV, and the conclusion explicitly says 'We conjecture that this statement holds to all n.' So the abstract presents a conjecture as a proven result. That is a real overstatement, though not a fatal one. The lower bound presumably follows from the different collinear pole structures of the color orderings, but no proof is given, and the stress-test note is correct to flag it.

A smaller caveat: the 'exhaustion by degree-two relations' is also only checked up to the same n and is conjectural beyond. The paper is transparent about this in the body, so it is acceptable, but the abstract should carry the same qualification.

Everything else checks out. The methods are standard, the counting is consistent, the computational claims are reproducible from the ancillary files, and the citation pattern looks fine. The extension over BBDJS/DM is real even if incremental: one new degree-two six-point relation, plus the general existence argument.

Who is this for? Anyone computing with all-plus or MHV amplitudes, and anyone interested in the algebraic structure of rational amplitudes. It deserves a serious referee, and I would send it to review. The referee should ask for the abstract and conclusion to be aligned with what is actually proven: the exact-count statement should be labeled a conjecture for general n. With that revision, I would be happy to see it published.

Referee Report

3 major / 5 minor

Summary. The paper studies linear relations (syzygies) among color-ordered one-loop all-plus Yang--Mills amplitudes and tree-level MHV amplitudes. The authors argue on general grounds that after a common spinor rescaling and multiplication by a common monomial, any such amplitude lies in the two-dimensional space spanned by 1 and the Levi--Civita pseudoscalar, which forces the existence of polynomial identities among any triplet of amplitudes (Sec. III). They develop two methods to find these relations: a numerical linear-algebra method over finite fields/rationals (Sec. IV) and a factorization-based syzygy computation (Secs. V--VI). They present explicit relations and generators for all-plus amplitudes through n=7 and for MHV through n=8 (Tables I--II), showing in these cases that the amplitudes reduce to exactly two independent ones. The paper also contains the observation that, in special representations, the six-point all-plus syzygies vanish under symmetrization (Sec. IX).

Significance. If the claims hold, the paper identifies a general mechanism behind polynomial relations among rational scattering amplitudes and provides a systematic way to construct them. The explicit low-degree generators, the cross-validated relation counts, and the two complementary computational approaches are useful and reproducible contributions. The claim that one-loop all-plus and tree-level MHV amplitudes reduce to exactly two independent color-ordered amplitudes is striking and goes beyond the KK/BCJ structure for MHV. The paper is also honest in providing exact expressions in auxiliary files and in cross-checking the numerical and analytic methods. However, the strongest advertised statement—the exact count for all n—is not established by the general theorem; it is verified only for small n and is explicitly conjectured in the conclusion.

major comments (3)

[Abstract and Sec. III (after Eq. 3.9)] The abstract states as a proven result that there are (n-1)!/2 - 2 relations for n≥5, “leaving only two independent color-ordered amplitudes.” The argument in Sec. III, however, constructs m-2 relations for m amplitudes, which shows that the span of the amplitudes has dimension at most two. It does not prove that the dimension is at least two. The exact count is verified only through n=7 for all-plus and n=8 for MHV (Tables I and II), and the Conclusion explicitly says “We conjecture that this statement holds to all n.” The abstract therefore overstates what is proven. The authors should either rephrase the claim as “at most two independent amplitudes” (with the exact count established numerically for the computed cases and conjectured in general), or supply the missing lower-bound argument (e.g., from distinct collinear pole structures of different orderings).
[Sec. IV and Tables I--II] The completeness of the relation counts rests on numerical rank computations. For the largest cases the paper uses finite-field arithmetic (K=F_p), and the authors state that they have exact rational expressions for the relations in all computed cases and cross-validate with the factorization method. This mitigates concerns, but the text should state explicitly which counts were obtained over Q and which over F_p, and should note that the equality of nullity over F_p and over Q is a reconstruction/verification issue. As written, the reader cannot tell from Tables I and II alone which entries rely solely on finite-field ranks.
[Sec. VIII and Conclusion] The claim that MHV tree amplitudes, after the KK and BCJ identities, are further reduced by degree-two relations to exactly two independent amplitudes is surprising: for n>5 this goes well beyond the standard (n-3)! count. The current evidence is the n≤8 computation in Table II, while the general theorem supplies only the upper bound of two. The conclusion's conjecture is therefore an essential part of the paper's central statement. I ask the authors to present this distinction clearly in the introduction and abstract, and to state explicitly that the exactness of the “two” for all n is conjectural unless a proof is added.

minor comments (5)

[Sec. III, Eq. (3.5)] The phase-removal factor as written, 1/([12][23]···[n1]) times the product of all [ij], is indeed polynomial because the denominator factors cancel against part of the numerator, but this is not immediately obvious in the displayed formula; a short clarification would help.
[Sec. IV, around Eq. (4.1)] The phrase “dim ker V = dim V − rank V” is a tautology; the intended statement is that the number of relations between the amplitudes equals dim V minus the rank of the sampled matrix. Rephrase to avoid confusion about what V is.
[Sec. V] The distinction between ordinary collinear kinematics and the “complex (or holomorphic) collinear kinematics” used in this work is central to the method, but it is introduced only briefly. A short explanatory paragraph or an example would make the section more accessible.
[Sec. IX] The statement that “it is also possible to find a non-minimal-degree representation where all syzygies vanish under symmetrization” is not demonstrated; pointing to a specific example in the auxiliary files would make the observation verifiable.
[Throughout] There are several minor typographical issues: in Eq. (4.3) the notation “I +” should likely be “I^+”, and in Sec. II the spacing in A(L)_{n;c} and related symbols is inconsistent. These do not affect the results.

Circularity Check

0 steps flagged · score 0.0 of 10

Derivation is self-contained: the claimed polynomial syzygies are constructed from the amplitudes' rational phase-weight structure, not assumed; no fitted parameter is renamed as a prediction.

full rationale

The central existence argument (Sec. III, Eqs. 3.7-3.9) starts from the explicit all-plus amplitude formula (3.2) and constructs, by multiplication by common spinor factors and clearing Gram-determinant denominators, polynomial expressions of the form Polya + Polyb epsilon_b. The minors xi_i in (3.8) then give explicit polynomial identities (3.9) for any triplet; this is a direct derivation, not an assumption of the target relations. The explicit low-degree syzygies for n=5,6,7 all-plus and n=6,7,8 MHV are obtained either by solving factorization equations (Sec. V) or by numerical linear algebra on the known amplitude expressions (Sec. IV); they are not obtained by fitting parameters to the desired relation count, and the counts are cross-checked. The only self-citations (refs. [8], [10]) are to standard color-decomposition and momentum-twistor parametrization tools; they are technical, not load-bearing, and no uniqueness theorem is imported from prior work of the authors. The paper itself states in the Conclusion that the exhaustive low-degree statement for all n is a conjecture ('We conjecture that this statement holds to all n'), and the abstract's exact-count phrasing is stronger than the proven 'at most two independent' bound from Sec. III; this is a completeness/correctness caveat, not a circularity.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The central claim rests on the rationality of the amplitudes and the 4D reduction of Lorentz invariants; these are stated domain assumptions. No free parameters are fitted to the target result, and no new physical entities are introduced. The explicit syzygy vectors are outputs of the computation, not inputs.

assumptions (5)

domain assumption The amplitudes are rational functions of the spinor variables with a common spinor phase weight and no branch cuts.
Invoked in Sec III to reduce each amplitude, after multiplication by a common monomial, to a polynomial of the form A(s) + B(s) ε_b. Holds for the all-plus one-loop and tree-level MHV amplitudes, but not for generic loop amplitudes with cuts.
domain assumption Four-dimensional kinematics: the space of contracted Levi-Civita tensors is spanned by a single ε_b.
Used in Sec III (around eq. 3.7) to write every expression as A(s) + B(s) ε_b; fails in d>4 dimensions or with dimensional regularization.
standard math The known explicit formulas for the all-plus and MHV amplitudes (e.g., eq. 3.2) are correct.
The amplitude formula (3.2) from Bern-Dixon-Kosower is the input for all computations; it is an external benchmark, not derived in this paper.
domain assumption Complex factorization channels for these amplitudes are only at ⟨i j⟩=0, with three-point amplitudes in the residues.
Used to derive the factorization equations (5.3)-(5.7) in Sec V; valid for all-plus and MHV but not for amplitudes with multiparticle poles.
domain assumption The numerical rank computations over finite fields at generic momentum-twistor points correctly compute the rank over the rationals.
Relied on for the completeness counts in Tables I and II and for verifying the candidates from the factorization method; standard in the field but not proven rigorously in the paper.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Serendipitous Syzygies of Scattering Amplitudes." pith.science (2026). https://pith.science/paper/WPD7L7BU

@misc{pith2026250514857,
  author       = {Pith},
  title        = {Pith review of: Serendipitous Syzygies of Scattering Amplitudes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WPD7L7BU}},
  note         = {Machine review of arXiv:2505.14857}
}

abstract

We study linear relations between color-ordered all-plus amplitudes at one loop in Yang--Mills theory. We show that on general grounds, there are $(n-1)!/2-2$ relations for $n\ge 5$, leaving only two independent color-ordered amplitudes. We present two complementary approaches to finding such relations: one using numerical linear algebra and the other using syzygies in computational algebraic geometry. We obtain explicit forms for all relations through $n=7$. We also study relations for the tree-level MHV amplitudes through $n=8$. The latter relations include the well-known color and Bern--Carrasco--Johansson identities.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

14 extracted references · 4 canonical work pages

[1]

Kleiss and H

R. Kleiss and H. Kuijf, Nucl. Phys. B312(1989), 616-644 doi:10.1016/0550-3213(89)90574-9; V. Del Duca, L. J. Dixon and F. Maltoni, Nucl. Phys. B571, 51-70 (2000) doi:10.1016/S0550- 3213(99)00809-3 [arXiv:hep-ph/9910563 [hep-ph]]

arXiv 1989
[2]

Z. Bern, J. J. M. Carrasco and H. Johansson, Phys. Rev. D78, 085011 (2008) doi:10.1103/PhysRevD.78.085011 [arXiv:0805.3993 [hep-ph]]

arXiv 2008
[3]

A. C. Edison and S. G. Naculich, JHEP09(2012), 069 doi:10.1007/JHEP09(2012)069 [arXiv:1207.5511 [hep-th]] A. C. Edison and S. G. Naculich, Nucl. Phys. B858, 488-501 (2012) doi:10.1016/j.nuclphysb.2012.01.019 [arXiv:1111.3821 [hep-th]] S. G. Naculich, Phys. Lett. B707, 191-197 (2012) doi:10.1016/j.physletb.2011.12.010 [arXiv:1110.1859 [hep-th]]

arXiv 2012
[4]

D. C. Dunbar, JHEP10, 058 (2023) doi:10.1007/JHEP10(2023)058 [arXiv:2308.06602 [hep- th]]

arXiv 2023
[5]

N. E. J. Bjerrum-Bohr, P. H. Damgaard, H. Johansson and T. Sondergaard, JHEP05(2011), 039 doi:10.1007/JHEP05(2011)039 [arXiv:1103.6190 [hep-th]]

work page Pith review arXiv 2011
[6]

L. J. Dixon and A. Morales, JHEP08, 129 (2024) doi:10.1007/JHEP08(2024)129 [arXiv:2407.13967 [hep-th]]; 22 L. J. Dixon and A. Morales, [arXiv:2411.10967 [hep-th]]

arXiv 2024
[7]

K. J. Costello, [arXiv:2302.00770 [hep-th]]

arXiv
[8]

Z. Bern, L. J. Dixon, D. C. Dunbar and D. A. Kosower, Nucl. Phys. B425, 217-260 (1994) doi:10.1016/0550-3213(94)90179-1 [arXiv:hep-ph/9403226 [hep-ph]]

arXiv 1994

Show all 14 references

[9]

Z. Bern, L. J. Dixon and D. A. Kosower, [arXiv:hep-th/9311026 [hep-th]]; G. Mahlon, Phys. Rev. D49, 4438-4453 (1994) doi:10.1103/PhysRevD.49.4438 [arXiv:hep- ph/9312276 [hep-ph]]; Z. Bern, G. Chalmers, L. J. Dixon and D. A. Kosower, Phys. Rev. Lett.72, 2134-2137 (1994) doi:10....

1994 arXiv
[10]

D. A. Kosower and S. Pögel, [arXiv:2206.14445 [hep-ph]]

arXiv
[11]

Dumas and P

J.-G. Dumas and P. Giorgi and C. Pernet, ACM Trans. on Mathematical Software (TOMS), 35, 3, 2008, 0098-3500, 1–42, 10.1145/1391989.1391992, ACM Press, New York, NY, USA; The FFLAS–FFPACK group, FFLAS–FFPACK: Finite Field Linear Algebra Subroutines / Package, v2.4.1, 2019,http:...

2008
[12]

Britto, F

R. Britto, F. Cachazo, B. Feng and E. Witten, Phys. Rev. Lett.94, 181602 (2005) doi:10.1103/PhysRevLett.94.181602 [arXiv:hep-th/0501052 [hep-th]]

2005 arXiv
[13]

Kawai, D

H. Kawai, D. C. Lewellen and S. H. H. Tye, Nucl. Phys. B269, 1-23 (1986) doi:10.1016/0550- 3213(86)90362-7

1986 doi
[14]

F. A. Berends, W. T. Giele and H. Kuijf, Phys. Lett. B211, 91-94 (1988) doi:10.1016/0370- 2693(88)90813-1; Z. Bern, D. C. Dunbar and T. Shimada, Phys. Lett. B312, 277-284 (1993) doi:10.1016/0370- 2693(93)91081-W [arXiv:hep-th/9307001 [hep-th]]; Z. Bern, L. J. Dixon, D. C. Dunb...

1988 arXiv

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

Kleiss and H

R. Kleiss and H. Kuijf, Nucl. Phys. B312(1989), 616-644 doi:10.1016/0550-3213(89)90574-9; V. Del Duca, L. J. Dixon and F. Maltoni, Nucl. Phys. B571, 51-70 (2000) doi:10.1016/S0550- 3213(99)00809-3 [arXiv:hep-ph/9910563 [hep-ph]]

arXiv 1989

[2] [2]

Z. Bern, J. J. M. Carrasco and H. Johansson, Phys. Rev. D78, 085011 (2008) doi:10.1103/PhysRevD.78.085011 [arXiv:0805.3993 [hep-ph]]

arXiv 2008

[3] [3]

A. C. Edison and S. G. Naculich, JHEP09(2012), 069 doi:10.1007/JHEP09(2012)069 [arXiv:1207.5511 [hep-th]] A. C. Edison and S. G. Naculich, Nucl. Phys. B858, 488-501 (2012) doi:10.1016/j.nuclphysb.2012.01.019 [arXiv:1111.3821 [hep-th]] S. G. Naculich, Phys. Lett. B707, 191-197 (2012) doi:10.1016/j.physletb.2011.12.010 [arXiv:1110.1859 [hep-th]]

arXiv 2012

[4] [4]

D. C. Dunbar, JHEP10, 058 (2023) doi:10.1007/JHEP10(2023)058 [arXiv:2308.06602 [hep- th]]

arXiv 2023

[5] [5]

N. E. J. Bjerrum-Bohr, P. H. Damgaard, H. Johansson and T. Sondergaard, JHEP05(2011), 039 doi:10.1007/JHEP05(2011)039 [arXiv:1103.6190 [hep-th]]

work page Pith review arXiv 2011

[6] [6]

L. J. Dixon and A. Morales, JHEP08, 129 (2024) doi:10.1007/JHEP08(2024)129 [arXiv:2407.13967 [hep-th]]; 22 L. J. Dixon and A. Morales, [arXiv:2411.10967 [hep-th]]

arXiv 2024

[7] [7]

K. J. Costello, [arXiv:2302.00770 [hep-th]]

arXiv

[8] [8]

Z. Bern, L. J. Dixon, D. C. Dunbar and D. A. Kosower, Nucl. Phys. B425, 217-260 (1994) doi:10.1016/0550-3213(94)90179-1 [arXiv:hep-ph/9403226 [hep-ph]]

arXiv 1994