Pith. sign in

REVIEW 4 major objections 5 minor 22 references

Representations of torsion-free arithmetic matroids

T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read This paper proves that every torsion-free arithmetic matroid of rank r has at most $m(E)^{r-1}$ essential representations up to equivalence, and gives an algorithm that constructs them all.

desk verdict A solid, genuinely useful enumeration algorithm for torsion-free arithmetic matroids, with a corrected typo and a reproducibility note needed before publication. read the letter →

arxiv 1908.04137 v1 pith:V42C7DIL submitted 2019-08-12 math.CO

classification math.CO MSC 05B3552C3505E4506A07
keywords arithmeticmatroidstorsion-freerepresentabilitytoricarrangementssignedHermitenormalformreductionposetoflayersCohen-Macaulay
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

Torsion-free arithmetic matroids package the combinatorics of finite lists of integer vectors: a matroid together with integer multiplicities that record the sizes of torsion subgroups in quotients by the listed vectors. This paper establishes that, up to equivalence, a rank-r torsion-free arithmetic matroid has at most $m(E)^{r-1}$ essential representations, meaning representations in a lattice of rank r, and gives an explicit algorithm that lists them all. The key move is a reduction that collapses any such matroid to a surjective one, where the whole ground set spans the lattice and a single forced representation exists; all original representations are then obtained by multiplying that reduced representation by integer matrices in Hermite normal form of determinant $m(E)$. As an application, the authors compute all representations of a small example and find toric arrangements whose layer and independence posets are not Cohen-Macaulay, disproving two conjectures.

What carries the argument

The two load-bearing constructions are the reduction $\overline{M}$ of a quasi-arithmetic matroid, defined by $$\overline{m}(X)=\frac{\gcd\{m(B)\mid B\text{ basis and }|X\cap B|=\operatorname{rk}(X)\}}{\gcd\{m(B)\mid B\text{ basis}\}},$$ with rank unchanged, and the signed Hermite normal form: the lexicographically smallest Hermite normal form obtainable from a matrix by changing column signs. The reduction extracts the torsion-free, surjective core of a representable matroid; the signed Hermite normal form canonicalizes matrices up to left multiplication by $\mathrm{GL}(r,\mathbb{Z})$ and column sign changes, via a polynomial-time algorithm whose inner loop uses the fact that the orbit of each entry has size at most 4.

What would settle it

A concrete falsifier is to search all rank-3 torsion-free arithmetic matroids with $m(E)=2$: finding one with more than $2^2=4$ essential representations up to equivalence would refute the main bound, and a brute-force search for any representable matroid whose reduction is not representable would refute Theorem 4.6.

Watch

Extended reading notes

Core claim

The paper's central claim is that reduction funnels representation theory: if $A$ is an essential representation of the reduced matroid $\overline{M}$, then every essential representation of $M$ is equivalent to $HA$ for some integer matrix $H$ in Hermite normal form with $\det(H)=m(E)$. Consequently a torsion-free arithmetic matroid of rank $r$ has at most $m(E)^{r-1}$ essential representations up to equivalence. The reduction operation, which keeps the rank and replaces subset multiplicities by ratios of gcds over bases, preserves representability, and the surjective reduced matroid has a unique essential representation, so multiplying by the admissible $H$'s exhausts all possibilities.

Load-bearing premise

The load-bearing premise is that no representation is lost when one passes to the reduced matroid, where the whole ground set spans the lattice: if the reduction omitted a representation, or if the forced representation of the reduced matroid were not unique, the list $HA$ would be incomplete.

Editorial extensions

If this is right

  • Any torsion-free arithmetic matroid of rank $r$ has at most $m(E)^{r-1}$ essential representations up to equivalence, a finite bound depending only on the multiplicity of the whole ground set and the rank.
  • Representability of a torsion-free arithmetic matroid is decidable by the paper's algorithm: failure of the sign-forcing step certifies non-representability, and success produces a candidate that is checked directly against all subset multiplicities.
  • The signed Hermite normal form solves the equivalence problem for representations in polynomial time, so listing non-equivalent representations is a matter of linear algebra rather than search over group elements.
  • For the displayed example, the 13 non-equivalent representations of one arithmetic matroid yield exactly 3 non-isomorphic posets of layers, while their arithmetic independence posets are pairwise isomorphic.
  • The same computations exhibit toric arrangements whose poset of layers and arithmetic independence poset have homology with $\mathbb{Z}_5$ summands, so these posets are not Cohen-Macaulay over fields of characteristic 5 and are not shellable.

Reading between the lines

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

  • Because the bound depends only on the single integer $m(E)$ and the rank, not on the finer collection of basis multiplicities, one could test whether a sharper bound holds in terms of the matroid's lattice of flats; the known small examples would be the first place to look.
  • If the arithmetic independence poset is indeed an invariant of the arithmetic matroid, as the paper's closing question asks, then it could be studied without choosing a representation; the 13-representation example is consistent with this, and a second example with two different independence posets would settle the question negatively.
  • The reduction operation is a canonical simplification with the strong gcd property, so it may be useful beyond representability, for instance in computing Tutte-like invariants or in testing whether other matroid invariants are invariant under reduction.
  • A census-style run of the published algorithm over all rank-3 torsion-free arithmetic matroids with small $m(E)$ could reveal whether the $m(E)^{r-1}$ bound is ever tight and, if so, on which matroids.
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.

Referee Report

4 major / 5 minor

Summary. The paper studies torsion-free arithmetic matroids and their representations as integer matrices (equivalently, central toric arrangements). It introduces the strong gcd property, a new operation called reduction, and a signed Hermite normal form. The main result (Theorem 5.4) states that every essential representation of a torsion-free arithmetic matroid M is equivalent to HA, where A is the unique essential representation of its reduction Mbar and H is a matrix in Hermite normal form with det(H)=m(E). This yields Corollary 5.5: a rank-r torsion-free arithmetic matroid has at most m(E)^(r-1) essential representations up to equivalence, together with an algorithm to list them. The paper also presents a polynomial-time signed Hermite normal form algorithm, a decomposition criterion, and, as applications, two computational counterexamples to conjectures that the poset of layers and the arithmetic independence poset of toric arrangements are shellable/Cohen-Macaulay.

Significance. If correct, the paper gives an effective classification of representations of torsion-free arithmetic matroids, improving the earlier bound m(E)^r of [Pag17] to m(E)^(r-1) and providing a concrete algorithm with a Sage implementation. The signed Hermite normal form is a useful new algorithmic tool. The two counterexamples in Section 8 disprove widely discussed conjectures and are therefore significant for the theory of toric arrangements. The paper contains detailed combinatorial proofs in Sections 3-6 and the main results are crisp and potentially influential. However, the central enumeration theorem rests on two external uniqueness results and on the reduction theorem, and the computational disproofs are not independently reproducible from the manuscript, so the significance is conditional on those points being secured.

major comments (4)
  1. [Theorem 5.4, proof] The determinant chain in the proof of Theorem 5.4 contains the line 'm(E) = det(H)·m(E) = det(H)', which is not correct as printed; it would force det(H)=1 or m(E)=0. The intended statement should be m(E) = det(H') = det(H), using mbar(E)=1. Since det(H)=m(E) is exactly the conclusion of the theorem, this needs to be fixed.
  2. [Theorem 4.6] The proof that the reduction Mbar is representable is only sketched. It relies on Corollary 3.6 (which itself depends on [DM13, Remark 3.1]) and on the identification of Mbar with the arithmetic matroid M' of the reduced vectors. The step where gcd{m(B)|B basis} = |T|·|K/G| and the subsequent conclusion m'(X)=mbar(X) for all X need to be written out in detail, because Theorem 4.6 is the bridge that makes the reduction from the general case to the surjective case valid.
  3. [Section 5.1] The uniqueness of the surjective representation is imported wholesale from [Pag17] (absolute values of entries) and [Len17b, Lemma 6] (signs of entries). These results are not stated or proved in the paper, yet the completeness of the enumeration in Theorem 5.4 and the bound in Corollary 5.5 depend on them. Please either state the exact propositions and give proofs (even condensed), or clearly mark this as a black-box dependence.
  4. [Section 8] The two disproofs of the shellability conjectures depend on computer calculations (the 13 representations and the homology groups (0,Z5,Z48) and (0,Z5,Z73)) that are not reproducible from the manuscript. No script, commit identifier of the Arithmat repository, or detailed verification is provided. Since this is the paper's advertised application, a reproducible computation (e.g., a small script and the exact matrices) should be included or referenced unambiguously.
minor comments (5)
  1. [Lemma 3.4] There is a typo: 'matorid' should be 'matroid'.
  2. [Definition 4.1] The reduction is denoted with a bar in the text, but the displayed definition writes M = (E, rk, m) without bars, which makes the notation confusing.
  3. [Corollary 5.5, proof] The product formula for the number of Hermite normal form matrices with given diagonal entries is printed in a garbled way; the intended expression \prod_i d_i^{r-i} should be typeset clearly.
  4. [Section 8] The homology notation (0,Z5,Z48) should specify whether this is reduced homology and in which degrees.
  5. [Proposition 6.9] The running time is stated as O(r^{\theta-1} n^2 (r^2+n)), but the displayed bound includes subsumed terms; please state explicitly the assumption that integer operations are O(1) and simplify the expression.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the derivation is a genuine reduction to the surjective case and to external uniqueness theorems, not a renaming of assumptions.

full rationale

The main derivation is not circular. The reduction Mbar is defined directly from gcds of basis multiplicities (Definition 4.1), and Theorem 4.6 proves, rather than assumes, that a representable M has representable Mbar by exhibiting the quotient of a representing lattice and using the strong gcd property to match multiplicities. Section 5.1 imports the uniqueness of the surjective representation from [Len17b] and [Pag17]; these are prior theorems with stated assumptions that do not include the paper's target classification, so they are independent support even though [Pag17] is self-cited. Theorem 5.4 then uses only that uniqueness and the Hermite normal form to express every essential representation of M as H A; the bound m(E)^(r-1) in Corollary 5.5 counts the possible H's and is not used as an input. The determinant line 'm(E) = det(H)·m(E) = det(H)' appears to contain a typo, but it is not an equation that defines the target by construction. The Section 8 counterexamples rely on software computations not fully reproducible from the manuscript, which is a verifiability concern rather than circularity.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

No free parameters are fitted to data; all definitions are exact. The paper leans on standard matroid and arithmetic matroid axioms and on cited theorems for the surjective case. No new physical or mathematical entities are postulated.

assumptions (3)
  • standard math Standard definitions and axioms of matroids and arithmetic matroids from [Oxl11], [DM13], and [BM14] are assumed without proof.
    These form the framework of the paper; no alternative model is considered.
  • domain assumption Cited results on representable torsion-free arithmetic matroids are correct and applicable: gcd property [DM13, Remark 3.1], uniqueness and sign forcing for the surjective case [Len17b, Pag17], and the orientability criterion [Pag18, Proposition 8.3].
    The algorithm in Section 5 depends on these results and does not re-derive them.
  • domain assumption The Arithmat/Sage implementation correctly computes posets and homology used in Section 8.
    The counterexamples rely on these computations; no independent machine check or commit hash is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Representations of torsion-free arithmetic matroids." pith.science (2026). https://pith.science/paper/V42C7DIL

@misc{pith2026190804137,
  author       = {Pith},
  title        = {Pith review of: Representations of torsion-free arithmetic matroids},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/V42C7DIL}},
  note         = {Machine review of arXiv:1908.04137}
}
read the original abstract

We study the representability problem for torsion-free arithmetic matroids. By using a new operation called "reduction" and a "signed Hermite normal form", we provide and implement an algorithm to compute all the representations, up to equivalence. As an application, we disprove two conjectures about the poset of layers and the independence poset of a toric arrangement.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

22 extracted references · 20 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION output.nonempty.mrnumber duplicate missing pop "" 'skip if duplicate empty 'pop " " swap * " " * write if FUNCTION fin.entry add.period write mrnumber output.nonempty.mrnumber newline INTEGERS nameptr namesleft numnames FUNCTION format.language language empty "" " (" language * ")" * if FUNCTION form...

  2. [2]

    Bj\"orner, M

    A. Bj\"orner, M. Las Vergnas, B. Sturmfels, N. White, and G. M. Ziegler, Oriented matroids, second ed., Encyclopedia of Mathematics and its Applications, vol. 46, Cambridge University Press, Cambridge, 1999. 1744046

  3. [3]

    Br\"and\'en and L

    P. Br\"and\'en and L. Moci, The multivariate arithmetic T utte polynomial , Transactions of the American Mathematical Society 366 (2014), no. 10, 5523--5540. 3240933

  4. [4]

    Callegaro and E

    F. Callegaro and E. Delucchi, The integer cohomology algebra of toric arrangements, Advances in Mathematics 313 (2017), 746--802. 3649237

  5. [5]

    Cohen, A course in computational algebraic number theory, Graduate Texts in Mathematics, vol

    H. Cohen, A course in computational algebraic number theory, Graduate Texts in Mathematics, vol. 138, Springer, 1993

  6. [6]

    De Concini and C

    C. De Concini and C. Procesi, Topics in hyperplane arrangements, polytopes and box-splines, Springer, 2010

  7. [7]

    D'Al \` and E

    A. D'Al \` and E. Delucchi, Stanley-Reisner rings for symmetric simplicial complexes, G -semimatroids and Abelian arrangements, arXiv preprint 1804.07366 (2018)

  8. [8]

    Shellability of posets of labeled partitions and arrangements defined by root systems

    E. Delucchi, N. Girard, and G. Paolini, Shellability of posets of labeled partitions and arrangements defined by root systems, arXiv preprint 1706.06360 (2017)

Show all 22 references
  1. [9]

    D'Adderio and L

    M. D'Adderio and L. Moci, Arithmetic matroids, the T utte polynomial and toric arrangements , Advances in Mathematics 232 (2013), 335--367

  2. [10]

    Lenz , Computing the poset of layers of a toric arrangement, arXiv preprint 1708.06646 (2017)

    M. Lenz , Computing the poset of layers of a toric arrangement, arXiv preprint 1708.06646 (2017)

  3. [11]

    , Representations of weakly multiplicative arithmetic matroids are unique , arXiv preprint 1704.08607 (2017)

  4. [12]

    , Stanley-Reisner rings for quasi-arithmetic matroids, arXiv preprint 1709.03834 (2017)

  5. [13]

    Le Gall, Powers of tensors and fast matrix multiplication, Proceedings of the 39th International Symposium on Symbolic and Algebraic Computation, ACM, 2014, pp

    F. Le Gall, Powers of tensors and fast matrix multiplication, Proceedings of the 39th International Symposium on Symbolic and Algebraic Computation, ACM, 2014, pp. 296--303

  6. [14]

    Martino, Face module for realizable Z -matroids , Contributions to Discrete Mathematics 13 (2018), no

    I. Martino, Face module for realizable Z -matroids , Contributions to Discrete Mathematics 13 (2018), no. 2, 74--87

  7. [15]

    Newman, Integral matrices, Pure and Applied Mathematics, vol

    M. Newman, Integral matrices, Pure and Applied Mathematics, vol. 45, Academic Press, 1972

  8. [16]

    Oxley, Matroid theory, second ed., Oxford Graduate Texts in Mathematics, vol

    J. Oxley, Matroid theory, second ed., Oxford Graduate Texts in Mathematics, vol. 21, Oxford University Press, Oxford, 2011. 2849819

  9. [17]

    Pagaria , Combinatorics of Toric Arrangements , arXiv preprint 1710.00409 (2017)

    R. Pagaria , Combinatorics of Toric Arrangements , arXiv preprint 1710.00409 (2017)

  10. [18]

    , Orientable arithmetic matroids , arXiv preprint 1805.11888 (2018)

  11. [19]

    , Two examples of toric arrangements, Journal of Combinatorial Theory, Series A 167 (2019), 389--402

  12. [20]

    Paolini, Shellability of generalized Dowling posets, arXiv preprint 1811.08403 (2018)

    G. Paolini, Shellability of generalized Dowling posets, arXiv preprint 1811.08403 (2018)

  13. [21]

    Pagaria and G

    R. Pagaria and G. Paolini, Arithmat: Sage implementation of arithmetic matroids and toric arrangements, https://github.com/giove91/arithmat, 2019

  14. [22]

    Storjohann and G

    A. Storjohann and G. Labahn, Asymptotically fast computation of Hermite normal forms of integer matrices, ISSAC 1996, 1996

Pith tools

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