Perfect matroids over hyperfields

Nathan Bowler; Rudi Pendavingh

REVIEW 3 major objections 4 minor 20 references

Perfect matroids over hyperfields

T0 review · 3 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read Over every stringent skew hyperfield, weak matroids are strong: all vectors and covectors are orthogonal.

desk verdict A genuine unification of perfection results for stringent skew hyperfields, but with an unproved bridging lemma and a heavy external classification that a referee should push on. read the letter →

arxiv 1908.03420 v3 pith:OBHJPB2C submitted 2019-08-09 math.CO

Nathan Bowler , Rudi Pendavingh This is my paper · ORCID

classification math.CO MSC 05B35

keywords stringenthyperfieldsmatroidsovervaluatedorientedvectoraxiomsresidueorthogonalityofcircuitsandcocircuitsskew

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

The paper establishes that stringent skew hyperfields—hyperfields in which addition is multivalued only when the two summands are negatives—form a class over which the weak and strong definitions of matroid coincide. The key theorem says that for any left H-matroid over such an H, every vector is orthogonal to every covector; from this the authors derive that H is perfect, hence weak matroids over H are strong matroids. The paper also gives a vector-axiom characterization: a subset of $H^E$ is the set of vectors of some left H-matroid exactly when it satisfies closure under scaling, a residue-dependent composition, and an elimination rule. This matters because valued fields, via the standard quotient construction, produce stringent hyperfields, so the framework simultaneously covers classical, oriented, valuated, and certain algebraic matroids while preserving the good minor and vector behavior that fails for general hyperfields.

What carries the argument

The load-bearing machinery is the structural classification of stringent skew hyperfields as semidirect products $H=R\rtimes_{U,\psi}\Gamma$: a residue hyperfield R equal to the Krasner hyperfield, the sign hyperfield, or a skew field, a group U mapped by $\psi$ onto an ordered value group $\Gamma$, with stable sums. From a matroid M over H, the paper builds the residue matroid $M_0$ over R whose circuits and cocircuits are the minimal supports $X^{\uparrow}$ of the original circuits and cocircuits, where $X^{\uparrow}$ records the entries of maximal size. The central transfer lemma says orthogonality passes down to these supports and, when two supports meet, back up again; this lets the authors lift the known perfection of the three residue types to all of H. Around this core sit a composition operation $\circ$ on H, used in the vector axiom (V2)$^1$, and minor theorems showing that vectors of contractions and deletions are exactly restrictions of vectors, which carry the inductions.

What would settle it

Exhibit a left H-matroid over a stringent skew hyperfield with a vector V and a covector U such that $0\notin\sum_{e\in E} V_e U_e$; Theorem 1 asserts no such pair exists. A finite search over small stringent hyperfields and small ground sets, checking orthogonality of all circuits against all cocircuits, would either find such a pair or confirm the claim in the tested range.

Watch

Extended reading notes

Core claim

Over a stringent skew hyperfield H, the paper proves Theorem 1: if M is a left H-matroid on a finite ground set E, V is a vector of M, and U is a covector of M, then V is orthogonal to U, meaning $0 \in \sum_{e\in E} V_e U_e$. Since the paper defines a hyperfield to be perfect exactly when this orthogonality holds for all matroids over it, this says every stringent skew hyperfield is perfect, and then Theorem 7 gives the equivalence of weak and strong matroids: every weak H-matroid has strong duality. The second main result, Theorem 44, characterizes vector sets: a set $\mathcal{V}\subseteq H^E$ is the vector set of a left H-matroid if and only if it satisfies (V0) $0\in\mathcal{V}$, (V1) closure under scalar multiplication, (V2)$^1$ closure under the composition operation whenever $V\circ W=V\cup W$, and (V3) the elimination rule that from $V_e=-W_e\neq 0$ one can find $Z\in\mathcal{V}$ with $Z_e=0$ and $Z\in V\oplus W$; the circuits are then exactly the minimal nonzero vectors.

Load-bearing premise

The proof assumes the imported classification theorem that every stringent skew hyperfield is a semidirect product whose residue is the Krasner hyperfield, the sign hyperfield, or a skew field; if that classification misses a case, the three-case reduction does not cover all stringent skew hyperfields.

Editorial extensions

If this is right

For every stringent skew hyperfield, weak matroids and strong matroids coincide: every weak H-matroid has globally orthogonal circuits and cocircuits, so the dual is well-behaved.
Vector sets of H-matroids are cryptomorphically described by the four vector axioms, so membership in a vector set can be checked by closure under scaling, composition, and elimination rather than by quantifying over bases.
Minors of H-matroids behave like minors of oriented and valuated matroids: contracting or deleting an element restricts the vectors exactly as expected, a property that fails for general hyperfields.
The residue matroid construction unifies the classical residue matroids of valuated matroids, oriented initial matroids, and residue linear spaces over skew fields within a single construction.
Because valued skew fields produce stringent hyperfields through the quotient construction, a linear space over a valued field carries its valuation and residue structure in one perfect matroid, so the two layers of data cannot contradict each other.

Reading between the lines

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

The three-case reduction suggests a testable boundary: if 'stringent' is relaxed while keeping the residue classification and support-transfer lemma intact, perfection may still hold; the paper's own examples show the minor and vector behavior fails for general hyperfields, so the boundary is likely near stringent.
The composition operation is associative only when the residue is the Krasner hyperfield and commutative only when it is not the sign hyperfield, so the vector axioms require only a weak, possibly non-associative composition; this hints that other axiom systems for matroids over tracts could be weakened in the same way.
For algebraic matroids in positive characteristic, the paper's field-extension construction packages Frobenius rescalings as rescalings of one matroid; a concrete next step would be to test whether the derivation space discussed in the field-extension section can be read off directly from the residue matroid, giving a local, characteristic-free criterion.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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

Here's the quick take: this is a real advance, and the main theorem is probably true, but there's one internal gap and one heavy external dependency that a referee needs to push on.

The paper proves that over any stringent skew hyperfield, every vector of a weak matroid is orthogonal to every covector, hence weak matroids are strong. That unifies earlier perfection results for doubly distributive hyperfields, oriented matroids, valuated matroids, and skew fields, and it extends them to noncommutative hyperfields. The vector axioms in Theorem 44 are a genuine generalization of the Murota–Tamura and oriented matroid axioms, and Theorem 40 shows that vector sets behave well under minors, which was not known for this class. The residue matroid construction, especially the noncommutative case in Lemma 31, is substantial and looks plausible. The paper is honest about the one place it cannot get a clean single-circuit elimination axiom when the residue is a skew field.

Now the soft spots, in proportion. Lemma 30 is the bridge that turns residue orthogonality into H-orthogonality, and it is stated without proof. The forward direction is easy; the converse, when the top supports intersect, needs an argument that uses the stable-sums condition from the Bowler–Su classification. This is not a trivial consequence of Lemma 13, and since Theorem 33 and Corollary 34 depend on it, a referee should insist on a proof. It may be fillable without changing the architecture, but it is a real missing piece.

The second dependency is the classification of stringent skew hyperfields [BS20], also by one of the present authors. Theorem 27 is imported wholesale. That is not a flaw per se—the classification is a separate paper—but the present paper's three-case proof collapses without it, so the referee should ask for a precise statement and, ideally, enough of the argument to see why the three residues are exhaustive.

Overall, I do not see a load-bearing error. The reductions are careful, the vector axioms are checked in both directions, and the limitations are acknowledged. The main theorem is significant and the paper deserves serious peer review. I'd ask for a proof of Lemma 30 and a clearer statement of the external classification before acceptance. It is a good paper for anyone working on matroid theory, tropical geometry, or hyperfields; I would bring it to a reading group and cite it.

Referee Report

3 major / 4 minor

Summary. The paper investigates matroids over stringent skew hyperfields, a class of hyperfields that includes the Krasner hyperfield, the sign hyperfield, skew fields, and the hyperfields arising from valued fields via Krasner's construction. The two central results are Theorem 1, which states that for any stringent skew hyperfield H and any left H-matroid M, every vector of M is orthogonal to every covector of M (so H is perfect), and Theorem 44, which gives vector axioms characterizing exactly the sets of vectors of left H-matroids over such H. The proofs proceed by using the Bowler--Su classification of stringent skew hyperfields to reduce to three cases according to the residue hyperfield (Krasner, sign, or skew field), constructing a residue matroid M0 for each H-matroid M, and then importing perfection and vector-axiom results from the known cases. The paper also proves that vectors behave well under taking minors for stringent hyperfields, generalizing results of Anderson and of Murota--Tamura.

Significance. If the main theorems hold, the paper unifies and extends a substantial body of work: it generalizes Dress--Wenzel's perfection results for fuzzy rings and the Baker--Bowler treatment of hyperfields, it provides a common framework for oriented matroids, valuated matroids, and matroids over skew fields, and it supplies vector axioms that specialize to the Murota--Tamura axioms for valuated matroids and to the standard oriented-matroid vector axioms. The applications sketched in Section 5, particularly to real tropical singularities and to algebraic matroids via derivations, indicate that the framework has real explanatory power. The manuscript is generally careful and detailed, with full proofs for most internally developed results, and the construction of the residue matroid is nontrivial and interesting. However, the present version has a load-bearing gap: Lemma 30, which bridges residue orthogonality to orthogonality over H, is stated without proof, and the main theorem depends on it. For this reason I cannot recommend acceptance in the current form.

major comments (3)

[§4.2, Lemma 30] Lemma 30 is stated without proof, but its converse direction is load-bearing for the paper's central claim. The lemma claims that for X,Y ∈ H^E with X^Ò ∩ Y^Ò ≠ ∅, residue orthogonality X^Ò ⟂ Y^Ò implies X ⟂ Y over H. This is not a formal consequence of the analogous Lemma 13 for Γ_max, because when the residue R is a skew field, orthogonality of the top parts only says that the residue sum is zero; to obtain 0 in the H-hypersum one must lift the residue terms with unit-group factors satisfying ψ(g_e)=1 and then absorb all lower-valuation contributions. That lifting argument is exactly where the stable-sums condition in the Bowler--Su description enters, but the paper never supplies it. The gap is inherited by Theorem 33 (the final step concludes V ⟂ U from V^Ò ⟂ U^Ò), by Corollary 34, and by Theorem 1. Please provide a complete proof of Lemma 30, or move it to an appendix with a full argument; the current one-line statement is not sufficient for such a central step.
[§4.1, Theorem 27] The main theorems depend essentially on the classification Theorem 27, imported from [BS20], which asserts that every stringent skew hyperfield has the form R ⋊_{U,ψ} Γ with R equal to the Krasner hyperfield, the sign hyperfield, or a skew field. This is an external result, and at the time of writing [BS20] is a preprint by one of the present authors and Ting Su. Since every reduction in Section 4, including the construction of the residue matroid and the three-case perfection argument, relies on this classification, the paper's main theorem is conditional on it. This is not a circularity, but it is a completeness concern: the reader cannot verify Theorem 1 from the present manuscript alone. The authors should either include a proof of the classification, state explicitly that the main theorem is a consequence of [BS20], or cite a published version if one now exists.
[§4.2, Lemma 32 and Theorem 33] The proof of Lemma 32 contains the sentence 'it follows that V M Y by Lemma 32', which cites the very lemma being proved; the intended reference is presumably Lemma 30. This is not merely a typo: it underscores that the missing proof of Lemma 30 is being used as a black box. Similarly, in the proof of Theorem 33, after the rescaling step the text says 'U^Ò ∈ UpMρ q', which should almost certainly be 'U^Ò ∈ UpM0q'; the notation Mρ is not defined at that point and the covector rescalings have not been tracked carefully. These points should be corrected along with the proof of Lemma 30.

minor comments (4)

[§2.4, Theorem 15 proof] In the proof of Theorem 15, the phrase 'we apply (D2) for C and D' appears to be a typo for '(M2)'.
[§4.3, Theorem 44 statement] The vector axioms in Theorem 44 refer to 'VpMq' inside the statement of the axioms before the matroid M has been constructed; these occurrences should be 'V' throughout, as is already done in the introductory Theorem 2.
[§3.1, Lemma 13] Lemma 13 is stated for Γ_max with a one-line proof relying on the order structure; it would help the reader to spell out the notation X^Ò · Y^Ò and the case distinction in the converse, since this lemma is the model for the unproved Lemma 30.
[§5.2] The name 'Jürgens' is garbled in the text and in the reference list; it should be spelled consistently, and the citation [J18] should give the full author name.

Circularity Check

1 steps flagged · score 2.0 of 10

The main derivation is not circular, but Lemma 32's proof contains a self-referential citation that, taken literally, invokes the lemma being proved.

other [Lemma 32 proof, Section 4.2 (residue matroid transfer)]
"Then V Ò MY Ò, and it follows that V MY by Lemma 32."

As written, the proof of Lemma 32 invokes Lemma 32 itself to pass from non-orthogonality of the top supports V^Ò and Y^Ò to non-orthogonality of V and Y. Taken literally, this proves the lemma by the lemma, a self-referential step. The evident intended citation is Lemma 30, which is exactly the unproved bridge statement 'if X^Ò ∩ Y^Ò ≠ ∅, then X^Ò ⟂ Y^Ò ⇒ X ⟂ Y'; if so, this is a citation typo and the real weakness is an omitted proof of Lemma 30, a correctness gap rather than a conceptual circularity. Either way, the step should be flagged.

full rationale

The central claims are that every stringent skew hyperfield is perfect (Corollary 34) and that the vector axioms characterize vector sets of left H-matroids (Theorem 44). The proof chain is: classify H as R ⋉_{U,ψ} Γ with R equal to the Krasner hyperfield, the sign hyperfield, or a skew field (Theorem 27, cited from Bowler–Su [BS20]); construct a residue matroid M0 over R (Lemma 31); transfer vectors and covectors to M0 via Lemma 32; invoke the known perfectness of K, S, and skew fields (Theorem 6); and lift orthogonality back to H via Lemma 30. None of these steps assumes the target statement 'H is perfect'; the conclusion appears only at the end. The Bowler–Su classification is self-adjacent because one of the present authors is a coauthor, but it is a parameter-free structural classification whose stated assumptions do not include the target result, so it counts as independent support and does not make the argument circular. Theorem 44 is a genuine cryptomorphism: sufficiency builds circuits from V, proves V = V(M) using Lemmas 35, 36, and 43, and necessity verifies the axioms from the vector set; no fitted parameter is renamed as a prediction and no known result is merely renamed. The only flagged item is the self-referential line in Lemma 32's proof, which is most plausibly a typo for Lemma 30; Lemma 30 itself is stated without proof, and that is a rigor/correctness issue rather than a circular reduction. Overall, the derivation is not equivalent to its inputs.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The central claims are proved by reduction to the residue hyperfield R using the Bowler-Su classification. The paper introduces no fitted constants and no postulates beyond standard matroid theory, the classification, and known perfection of the three residue types. The composition operation and residue matroid are constructions within existing theory, not newly postulated primitive entities.

assumptions (4)

domain assumption Every stringent skew hyperfield is isomorphic to R semidirect product U,psi Gamma, where R is the Krasner hyperfield, the sign hyperfield, or a skew field (Bowler-Su classification, Theorem 27).
Used throughout Section 4 to reduce proofs about H to proofs about its residue R; without this classification the main perfection theorem and vector axioms lack their three-case foundation.
domain assumption The ground set E is finite and all matroids are finite matroids.
Stated in Theorems 2, 24, and 44, and used in induction proofs on |E| (for example, Lemmas 16, 31, and 41).
standard math The Krasner hyperfield, the sign hyperfield, and skew fields are perfect (Theorem 6).
The base cases for Theorem 33; the paper cites known results for oriented matroids and linear spaces but does not prove them here.
standard math Minty's matroid characterization (Theorem 14) is valid.
Used in the proof of Lemma 16 and Theorem 15 to construct residue matroids from support sets.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Perfect matroids over hyperfields." pith.science (2026). https://pith.science/paper/OBHJPB2C

@misc{pith2026190803420,
  author       = {Pith},
  title        = {Pith review of: Perfect matroids over hyperfields},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OBHJPB2C}},
  note         = {Machine review of arXiv:1908.03420}
}

abstract

We investigate valuated matroids with an additional algebraic structure on their residue matroids. We encode the structure in terms of representability over stringent hyperfields. A hyperfield $H$ is {\em stringent} if $a\boxplus b$ is a singleton unless $a=-b$, for all $a,b\in H$. By a construction of Marc Krasner, each valued field gives rise to a stringent hyperfield. We show that if $H$ is a stringent skew hyperfield, then the vectors of any weak matroid over $H$ are orthogonal to its covectors, and we deduce that weak matroids over $H$ are strong matroids over $H$. Also, we present vector axioms for matroids over stringent skew hyperfields which generalize the vector axioms for oriented matroids and valuated matroids.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

20 extracted references · 15 canonical work pages

[1]

Vectors of matroids over tracts

Laura Anderson. Vectors of matroids over tracts. J. Combin. Theory Ser. A , 161:236--270, 2019

2019
[2]

Matroids over partial hyperstructures

Matt Baker and Nathan Bowler. Matroids over partial hyperstructures. Preprint, available on arXiv:1709.09707, 2017

work page Pith review arXiv 2017
[3]

Bollen, Jan Draisma, and Rudi Pendavingh

Guus P. Bollen, Jan Draisma, and Rudi Pendavingh. Algebraic matroids and F robenius flocks. Adv. Math. , 323:688--719, 2018

work page 2018
[4]

o rner, Michel Las Vergnas, Bernd Sturmfels, Neil White, and G\

Anders Bj\" o rner, Michel Las Vergnas, Bernd Sturmfels, Neil White, and G\" u nter M. Ziegler. Oriented matroids , volume 46 of Encyclopedia of Mathematics and its Applications . Cambridge University Press, Cambridge, second edition, 1999

1999
[5]

Matroids and their Dressians

Madeline Brandt. Matroids and their D ressians, 2019. Preprint, available on arXiv:1902.05592

work page Pith review arXiv 2019
[6]

Classification of doubly distributive skew hyperfields and stringent hypergroups

Nathan Bowler and Ting Su. Classification of doubly distributive skew hyperfields and stringent hypergroups. Preprint, available on arXiv:2003.03751, 2020

work page Pith review arXiv 2003
[7]

Andreas W. M. Dress. Duality theory for finite and infinite matroids with coefficients. Adv. in Math. , 59(2):97--123, 1986

work page 1986
[8]

Andreas W. M. Dress and Walter Wenzel. Perfect matroids. Adv. Math. , 91(2):158--208, 1992

work page 1992

Show all 20 references

[9]

Andreas W. M. Dress and Walter Wenzel. Valuated matroids. Adv. Math. , 93(2):214--250, 1992

1992
[10]

The diagonal of tropical matroid varieties and cycle intersections

Georges Fran c ois and Johannes Rau. The diagonal of tropical matroid varieties and cycle intersections. Collectanea Mathematica , 64(2):185--210, 2013

2013
[11]

On the relation between hyperrings and fuzzy rings

Jeffrey Giansiracusa, Jaiung Jun, and Oliver Lorscheid. On the relation between hyperrings and fuzzy rings. Beitr. Algebra Geom. , 58(4):735--764, 2017

2017
[12]

Real tropical singularities and bergman fans, 2018

Christian Jürgens. Real tropical singularities and bergman fans, 2018. Preprint, available on arXiv:1802.01838

2018 arXiv
[13]

Approximation des corps valu\' e s complets de caract\' e ristique p =0 par ceux de caract\' e ristique 0

Marc Krasner. Approximation des corps valu\' e s complets de caract\' e ristique p =0 par ceux de caract\' e ristique 0 . In Colloque d'alg\`ebre sup\' e rieure, tenu \`a B ruxelles du 19 au 22 d\' e cembre 1956 , Centre Belge de Recherches Math\' e matiques, pages 129--206. \...

1956
[14]

A class of hyperrings and hyperfields

Marc Krasner. A class of hyperrings and hyperfields. Internat. J. Math. Math. Sci. , 6(2):307--311, 1983

1983
[15]

George J. Minty. On the axiomatic foundations of the theories of directed linear graphs, electrical networks and network-programming. J. Math. Mech. , 15:485--520, 1966

1966
[16]

On circuit valuation of matroids

Kazuo Murota and Akihisa Tamura. On circuit valuation of matroids. Adv. in Appl. Math. , 26(3):192--225, 2001

2001
[17]

Field extensions, derivations, and matroids over skew hyperfields

Rudi Pendavingh. Field extensions, derivations, and matroids over skew hyperfields. Preprint, available on arXiv:1802.02447, 2018

2018 arXiv
[18]

Computing tropical linear spaces

Felipe Rincón. Computing tropical linear spaces. Journal of Symbolic Computation , 51:86 -- 98, 2013. Effective Methods in Algebraic Geometry

2013
[19]

A tropical intersection product in matroidal fans

Kristin M Shaw. A tropical intersection product in matroidal fans. SIAM Journal on Discrete Mathematics , 27(1):459--491, 2013

2013
[20]

David E. Speyer. Tropical linear spaces. SIAM J. Discrete Math. , 22(4):1527--1558, 2008

2008

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Vectors of matroids over tracts

Laura Anderson. Vectors of matroids over tracts. J. Combin. Theory Ser. A , 161:236--270, 2019

2019

[2] [2]

Matroids over partial hyperstructures

Matt Baker and Nathan Bowler. Matroids over partial hyperstructures. Preprint, available on arXiv:1709.09707, 2017

work page Pith review arXiv 2017

[3] [3]

Bollen, Jan Draisma, and Rudi Pendavingh

Guus P. Bollen, Jan Draisma, and Rudi Pendavingh. Algebraic matroids and F robenius flocks. Adv. Math. , 323:688--719, 2018

work page 2018

[4] [4]

o rner, Michel Las Vergnas, Bernd Sturmfels, Neil White, and G\

Anders Bj\" o rner, Michel Las Vergnas, Bernd Sturmfels, Neil White, and G\" u nter M. Ziegler. Oriented matroids , volume 46 of Encyclopedia of Mathematics and its Applications . Cambridge University Press, Cambridge, second edition, 1999

1999

[5] [5]

Matroids and their Dressians

Madeline Brandt. Matroids and their D ressians, 2019. Preprint, available on arXiv:1902.05592

work page Pith review arXiv 2019

[6] [6]

Classification of doubly distributive skew hyperfields and stringent hypergroups

Nathan Bowler and Ting Su. Classification of doubly distributive skew hyperfields and stringent hypergroups. Preprint, available on arXiv:2003.03751, 2020

work page Pith review arXiv 2003

[7] [7]

Andreas W. M. Dress. Duality theory for finite and infinite matroids with coefficients. Adv. in Math. , 59(2):97--123, 1986

work page 1986

[8] [8]

Andreas W. M. Dress and Walter Wenzel. Perfect matroids. Adv. Math. , 91(2):158--208, 1992

work page 1992