Topological Koszulity for Category Algebras

David Favero; Pouya Layeghi

REVIEW 4 major objections 4 minor 8 references

Topological Koszulity for Category Algebras

T0 review · 4 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read The Koszulity of a category algebra is equivalent to a topological bouquet condition on its factorization spaces.

desk verdict Strong framework, but Proposition 2.40 is false for length-zero morphisms, and the main theorem's proof needs repair before the paper is correct as stated. read the letter →

arxiv 2412.18206 v2 pith:XSH7EYXL submitted 2024-12-24 math.RA math.AG

David Favero , Pouya Layeghi This is my paper · ORCID

classification math.RAmath.AG MSC 16S3716E3018G3506A11

keywords Koszulalgebrascategorylocallybouquetfactorizationspacesreducednerveincidencehomotopypathexceptionalcollections

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 proves that for a large class of graded categories, the Koszul property of the category algebra is exactly a topological property of the category. The topological invariant is a family of spaces BC(p), one for each morphism p, built from the reduced nerve of the nontrivial factorizations of p. The main theorem says kC is Koszul if and only if each BC(p) is a bouquet, meaning it has no reduced cohomology below its dimension. This unifies and recovers the classical theorem that incidence algebras of graded posets are Koszul exactly when the poset is locally Cohen-Macaulay, and it extends to homotopy path algebras and to endomorphism algebras of line bundles on toric varieties.

What carries the argument

The load-bearing object is the reduced nerve \bar N(C) of an indiscretely based category and, for each morphism p, the factorization space BC(p): the geometric realization of the semi-simplicial set of nontrivial factorizations p = f_0 \circ \dots \circ f_{n+1} modulo the relation that inserts isomorphisms between adjacent factors. This space carries a cellular projective resolution of the diagonal bimodule, and the complex computing Ext is the CW cohomology chain complex of BC(p). The identity doing the work is the formula expressing Ext^i(S_w,S_v)_{-n} as a direct sum of reduced cohomology groups \tilde $H^{{i-2}}$(BC(p)) over all paths of length n from v to w.

What would settle it

Take the entrance path category of the two-dimensional torus with the four-strata cube stratification from Example 4.29/4.40; the interval (~v,~w) is two disjoint arcs, so BC(p) for the corresponding path is disconnected and \tilde $H^{0}$(BC(p)) \neq 0. The theorem predicts $Ext^{2}$(S_w,S_v)_{-n} \neq 0 for some n \neq 2, hence non-Koszulity; a direct computation of $Ext^{2}$ from the quiver with the single cubic relation would settle whether the prediction holds.

Watch

Extended reading notes

Core claim

The central discovery is a topological formula for Ext groups between simple representations of an indiscretely based category: Ext^i_{kC}(S_w,S_v)_{-n} = \bigoplus_{p: l(p)=n, t(p)=v, h(p)=w} \tilde $H^{{i-2}}$(BC(p)), where BC(p) is the geometric realization of the factorization space of p. Consequently Koszulity, which by the standard criterion is the vanishing of Ext^i(A_0,A_0)_{-n} for i \neq n, becomes the statement that every BC(p) is cohomologically bouquet. The argument runs through a cellular projective resolution of the diagonal bimodule whose k-cells are indexed by cells of the reduced nerve, obtained from the normalized standard resolution after passing to a skeletal category. This reduces a purely algebraic property to a collection of topological vanishing statements.

Load-bearing premise

The entire equivalence assumes the category is indiscretely based: each connected component of the degree-zero morphisms has exactly one morphism between any two objects, so no nontrivial automorphisms exist.

Editorial extensions

If this is right

For every graded poset P, the incidence algebra kP is Koszul if and only if P is locally Cohen-Macaulay, recovering the known theorem as a corollary.
Almost discrete fibrations preserve Koszulity; Reiner-Stamate equivalence relations on graded posets are exactly such fibrations, so Koszulity of an incidence algebra and its reduced incidence algebra coincide.
A graded homotopy path algebra is Koszul if and only if its path poset is locally Cohen-Macaulay, turning a homological property into a purely combinatorial one.
For Bondal-Thomsen homotopy path algebras, Koszulity is equivalent to the vanishing of reduced cohomology of certain unions of strata associated to open intervals in the induced stratification.
For a full strong exceptional collection of line bundles on a toric variety with a graded ordering, Koszulity is equivalent to the existence of shifts making the shifted dual exceptional collection strong.

Reading between the lines

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

The Ext formula suggests that Koszulity could be tested algorithmically by computing reduced cohomology of finite semi-simplicial sets, which is more tractable than resolving the algebra directly.
The paper explicitly notes that nontrivial automorphisms would bring representation theory of automorphism groups into the story; a plausible extension would replace the simple functors by simples with equivariant structure and formulate a twisted, character-dependent bouquet condition.
Because Koszulity passes to saturated subcategories, the theorem gives a quick obstruction: any non-locally-bouquet subcategory inside a Koszul category would force non-Koszulity, a shortcut for identifying examples like the Hirzebruch surface F1.
The toric application indicates that strongness of a shifted dual exceptional collection can be certified by local Cohen-Macaulayness of monomial posets, which may be easier to check than computing all derived Hom spaces.

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

Worth a look, but the main computational proposition has a bug that needs fixing before the main theorem can be trusted. Proposition 2.40 claims a direct sum over all p with l(p)=n of reduced cohomology of BC(p). For n=0, p is an isomorphism, and BC(p) is empty. For i=1 that gives \tilde H^{-1}(∅)=k, but Ext^1(k[x])(k,k)_0=0. The same issue hits i=0,n=0, where the RHS is 0 but Hom is k. The proof via Lemma 2.39 only covers non-isomorphism p and cells of dimension at least 2, so the statement overreaches. This is not cosmetic: Theorem 2.66 invokes the equality for all i,n, and Proposition 2.62 inherits the problem. The fix is straightforward: restrict the sum to p of positive length (or non-isomorphisms) and handle Ext^0 and Ext^1 separately. I expect the theorem survives, but the paper as written is not correct.

What is genuinely new and good: the idea of using the reduced nerve and factorization spaces to get a topological handle on Ext between simples is a real step beyond the poset cases; the locally bouquet criterion is a natural common generalization of Reisner-style conditions for incidence algebras and the sufficient condition in FH22; and the almost discrete fibration transfer (Corollary 2.78) gives a clean functorial mechanism, which then produces the Reiner–Stamate correspondence in Section 3. The applications to Bondal–Thomsen HPAs and toric line bundle collections are also useful, and the paper is honest about relying on prior work.

Other soft spots: Proposition 2.62(3) states an iff that is only an implication (a length-one morphism is indecomposable, but the converse needs the previous condition), and the proof of Proposition 4.49 is compressed—it leans on star-shapedness and a homeomorphism claim that is plausible but not fully shown. These are minor compared to the Prop 2.40 issue.

Who is this for? People working on Koszul algebras, incidence algebras, and homotopy path algebras will want to read it, but they should treat the current version as a draft. It deserves a serious referee, not a desk reject, because the framework and applications are valuable and the flaw is local and fixable. I would not cite it in its present form; once corrected, likely yes.

Referee Report

4 major / 4 minor

Summary. The paper studies category algebras of N-graded categories that are "indiscretely based" (no nontrivial automorphisms; the degree-zero connected components are indiscrete categories). It introduces a reduced nerve BC and factorization spaces BC(p) and claims a formula (Proposition 2.40) expressing Ext^i between simple functors as a direct sum of reduced cohomology groups \tilde H^{i-2}(BC(p)). From this it derives the main theorem (Theorem 2.66): kC is Koszul if and only if C is locally bouquet. It then defines almost discrete fibrations, proves that Reiner–Stamate equivalence relations on posets are exactly such fibrations, recovers the Reiner–Stamate Koszulity criterion, and applies the framework to homotopy path algebras, Bondal–Thomsen algebras, and full strong exceptional collections of line bundles on toric varieties.

Significance. The conceptual package is valuable: reindexing the normalized standard resolution by cells of a reduced nerve is a natural idea, and the paper demonstrates its use by recovering the poset case, the Reiner–Stamate reduced incidence algebra theorem, and the Favero–Huang sufficient condition, while also giving new toric consequences. If the main theorem is correct, it provides a clean topological criterion for Koszulity in a broad class of algebras. However, the central computational proposition has a low-degree defect, and the main theorem is therefore not proved as written. The paper is worth pursuing after a focused revision that corrects the statement and proof of Proposition 2.40 and the statements that depend on it.

major comments (4)

[Section 2.3, Proposition 2.40] The displayed equality is false as stated for length-zero morphisms and for i=0,1. Take C with one object v and morphisms {x^n:n≥0}, so kC≅k[x] (Example 2.6). For p=id_v, BC(p)=∅ because there are no nontrivial factorizations of an isomorphism. With the convention \tilde H^{-1}(∅)=k adopted in the preamble to Proposition 2.62, the right-hand side for (i,n)=(1,0) contains the summand k from p=id_v, but Ext^1_{k[x]}(k,k)_0=0, as read off from the standard resolution 0→k[x](-1)→k[x]→k→0. For (i,n)=(0,0) the right-hand side is \tilde H^{-2}(∅)=0, while the left-hand side is Hom_{kC}(S_v,S_v)=k. Thus Proposition 2.40 cannot be used for all i,n as Theorem 2.66, Corollary 2.44, and Proposition 2.62 do. The proof appears to require restricting the direct sum to non-isomorphism morphisms and treating the degree-zero Ext^0 and Ext^1 terms separately; as written, the proposition is false.
[Section 2.4.4, Proposition 2.62] Condition (3), that a morphism is indecomposable if and only if it has length 1, is false for length-zero morphisms: every isomorphism, in particular every identity, is indecomposable because a nontrivial factorization cannot compose to an isomorphism. Moreover condition (2) cannot be interpreted as including p=id, since then \tilde H^{-1}(BC(id))=\tilde H^{-1}(∅)=k for every object, so (2) would fail for every category and would force (1) to fail, contradicting standard examples such as k[x]. The proposition and its proof need to be limited to positive-length morphisms and to exclude the identity from the vanishing condition. Proposition 2.64 inherits the same problem in its hypothesis that BC(p) is non-empty for all p with l(p)≠1.
[Section 2.4.4, proof of Theorem 2.66] The step that a direct sum of \tilde H^{i-2}(BC(p)) over all p of length n vanishes if and only if each BC(p) is bouquet is compressed and needs an explicit dimension argument. For a morphism p with l(p)=n, any nontrivial factorization has each factor a non-isomorphism, hence positive length, so a cell of dimension r in BC(p) corresponds to r+2 factors and therefore r+2≤n; thus dim BC(p)≤n-2. Reduced cohomology in degrees above the dimension vanishes automatically, so the vanishing required for Koszulity, namely \tilde H^{i-2}(BC(p))=0 for i≠n, is equivalent to \tilde H^j(BC(p))=0 for j<dim BC(p) in the relevant range. This should be stated explicitly; the current one-line equivalence hides the point and, together with the Proposition 2.40 issue, makes the proof of Theorem 2.66 incomplete.
[Section 2.1, Proposition 2.16] Proposition 2.16 is stated without the indiscretely based hypothesis and is false in general. If C is a one-object category whose morphisms form a nontrivial finite group G, then S_v is the trivial k[G]-module, which is not simple when the characteristic of k divides |G|, even though every morphism of C is an isomorphism. The proof implicitly uses the uniqueness of isomorphisms that is supplied later by the indiscretely based hypothesis. The proposition should be stated under that hypothesis or made conditional on it.

minor comments (4)

[Section 2.4.4, proof of Proposition 2.62] The symbol Q0 is used without definition; it should be Ob(C).
[Theorem 1.4 and Theorem 4.53] The notation \tilde H^i(Rn(D,E)) is missing the subscript on R^n; it should refer to R^n_{(D,E)} as defined in Section 4.2.3.
[Example 2.22] The phrase "they are in discretely based" appears to be a typo for "indiscretely based".
[Section 3, opening paragraph] The name Reiner is misspelled as "Riener" in the opening sentence.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the main Koszulity criterion is derived from Mitchell's resolution and BGS's Koszul criterion, not from its own target.

full rationale

The paper's central chain is non-circular. Theorem 2.66 derives Koszulity from Mitchell's normalized standard resolution (Mit72) and the Beilinson–Ginzburg–Soergel vanishing characterization of Koszulity (BGS96, Prop. 2.1.3). Proposition 2.36 builds the cellular bimodule resolution from the reduced nerve by passing through a skeletal category; Proposition 2.40 identifies the resulting Ext computation with the CW cohomology of factorization spaces. The property 'locally bouquet' (Defs. 2.53 and 2.56) is a topological vanishing condition, not defined as 'Ext vanishes', so the equivalence is contentful rather than definitional. Citations to FH22 are used for background and applications, not to justify the main algebraic theorem; Reiner–Stamate and earlier poset results are recovered as corollaries rather than assumed. A possible defect in Proposition 2.40 concerning length-zero morphisms and the convention ~H^{-1}(∅)=k is a correctness issue in the stated formula, not a circularity: no prediction is forced by a fitted input, and the target result is not assumed as an input.

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

The central proof relies on standard categorical and homological background (Mitchell-Webb equivalence, BGS Koszulity criteria) and on external results for the applications (Bondal, FH22). None of these are fitted to the target conclusion; they are inputs from the literature. The paper introduces no free parameters and no invented physical or geometric entities in the graviton sense.

assumptions (5)

standard math Mitchell-Webb equivalence: the category of representations of C is equivalent to the category of modules over the category algebra RC (Theorem 2.9).
Used throughout to compute Ext groups in the functor category rather than in the module category, after Mitchell 1972 and Webb 2007.
standard math Beilinson-Ginzburg-Soergel characterization of Koszulity via vanishing of Ext^i(A0,A0) for i != n (Proposition 2.47).
This external theorem is the bridge between the algebraic definition of Koszulity and the topological vanishing conditions.
domain assumption Bondal's result that Koszulity of the endomorphism algebra of a full strong exceptional collection is equivalent to strongness of the shifted dual collection (Bon90, Corollary 7.2).
Used in Proposition 4.63 and Theorem 4.64 to translate Koszulity into a statement about dual exceptional collections in toric geometry.
domain assumption Favero-Huang results on homotopy path algebras, including the identification of HPAs as category algebras and the simple-stratification theorem (FH22, Theorem 4.44, Corollary 5.11).
Used for the applications to stratified spaces and toric varieties; these are prior results of the same first author, not assumptions tailored to this paper's central claim.
domain assumption Standard facts about toric varieties: the Cox ring is the total coordinate ring, and the endomorphism algebra of a sum of line bundles is a skew category algebra kC_S.
Invoked in Example 4.9 and Proposition 4.59 without proof; standard in toric geometry, though not explicitly cited for this precise statement.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Topological Koszulity for Category Algebras." pith.science (2026). https://pith.science/paper/XSH7EYXL

@misc{pith2026241218206,
  author       = {Pith},
  title        = {Pith review of: Topological Koszulity for Category Algebras},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XSH7EYXL}},
  note         = {Machine review of arXiv:2412.18206}
}

read the original abstract

We give a topological description of Ext groups between simple representations of categories via a nerve type construction. We use it to show that the Koszulity of indiscretely based category algebras is equivalent to the locally bouquet property of this nerve. We also provide a class of functors which preserve the Koszulity of category algebras called almost discrete fibrations. Specializing from categories to posets, we show that the equivalence relations of V. Reiner and D. Stamate in arXiv:0904.1683 [math.AC] are exactly almost discrete fibrations and recover their results. As an application, we classify when a shifted dual collection to a full strong exceptional collection of line bundles on a toric variety is strong.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

8 extracted references · 5 canonical work pages

[1]

Cohen-Macaulay ordered sets

[Bac80] Kenneth Baclawski. “Cohen-Macaulay ordered sets” . In: Journal of Alge- bra 63.1 (1980), pp. 226–258. [BGS96] Alexander Beilinson, Victor Ginzburg, and Wolfgan g Soergel. “Koszul du- ality patterns in representation theory”. In: Journal of the American Math- ematical Society 9.2 (1996), pp. 473–527. [Bjö80] Anders Björner. “Shellable and Cohen-Mac...

work page 1980
[6]

On the cohomology of incidence alge- bras of partially ordered sets

56 REFERENCES [IZ90] Kiyoshi Igusa and Dan Zacharia. “On the cohomology of incidence alge- bras of partially ordered sets”. In: Communications in Algebra 18.3 (1990), pp. 873–887. [LR20] Fosco Loregian and Emily Riehl. “Categorical notion s of ﬁbration”. In: Expositiones Mathematicae 38.4 (2020), pp. 496–514. [Lur17] Jacob Lurie. Higher Algebra

work page 1990
[12]

2024, e24

Cambridge University Press. 2024, e24. [Huy06] Daniel Huybrechts. Fourier-Mukai transforms in algebraic geometry. Claren- don Press,

work page 2024
[183]

Posets, regular CW complexes and B ruhat order

[Bjö84] Anders Björner. “Posets, regular CW complexes and B ruhat order.” In: Eur. J. Comb. 5.1 (1984), pp. 7–16. [BK05] Eric Babson and Dmitry N Kozlov. “Group actions on pos ets”. In: Journal of Algebra 285.2 (2005), pp. 439–450. [Bon06] Alexey Bondal. Derived categories of toric varieties, Convex and algebrai c geometry, Oberwolfach conference reports, vol. 3

work page 1984
[1994]

Lectures on noncommutative resol utions

[Wem12] Michael Wemyss. “Lectures on noncommutative resol utions”. In: arXiv preprint arXiv:1210.2564 (2012). [Woo98] David Woodcock. “Cohen–Macaulay complexes and Kos zul rings”. In: Journal of the London Mathematical Society 57.2 (1998), pp. 398–410. [Yan05] Kohji Yanagawa. “Dualizing complex of the incidenc e algebra of a ﬁnite regular cell complex”. I...

arXiv 2012
[2006]

Helices, representations of quiv ers and Koszul algebras

[Bon90] Alexei I Bondal. “Helices, representations of quiv ers and Koszul algebras”. In: Helices and vector bundles 148 (1990), pp. 75–95. [BWW05] Anders Björner, Michelle Wachs, and Volkmar Welker . “Poset ﬁber the- orems”. In: Transactions of the American Mathematical Society 357.5 (2005), pp. 1877–1899. [CLS24] David A Cox, John B Little, and Henry K S...

work page 1990
[2017]

Rings with several objects

[Mit72] Barry Mitchell. “Rings with several objects”. In: Advances in Mathematics 8.1 (1972), pp. 1–161. [Pol95] Patrick Polo. “On Cohen-Macaulay posets, Koszul al gebras and certain modules associated to Schubert varieties”. In: Bulletin of the London Mathematical Society 27.5 (1995), pp. 425–434. [Pol97] Alexander Polishchuk. “Perverse sheaves on a tria...

arXiv 1972
[2024]

Moduli of McKay quiver representations I: the coherent component

[CMT07a] Alastair Craw, Diane Maclagan, and Rekha R Thomas. “Moduli of McKay quiver representations I: the coherent component”. In: Proceedings of the London Mathematical Society 95.1 (2007), pp. 179–198. [CMT07b] Alastair Craw, Diane Maclagan, and Rekha R Thomas. “Moduli of McKay quiver representations II: Gröbner basis techniques”. In: Journal of Alge- ...

arXiv 2007

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [1]

Cohen-Macaulay ordered sets

[Bac80] Kenneth Baclawski. “Cohen-Macaulay ordered sets” . In: Journal of Alge- bra 63.1 (1980), pp. 226–258. [BGS96] Alexander Beilinson, Victor Ginzburg, and Wolfgan g Soergel. “Koszul du- ality patterns in representation theory”. In: Journal of the American Math- ematical Society 9.2 (1996), pp. 473–527. [Bjö80] Anders Björner. “Shellable and Cohen-Mac...

work page 1980

[2] [6]

On the cohomology of incidence alge- bras of partially ordered sets

56 REFERENCES [IZ90] Kiyoshi Igusa and Dan Zacharia. “On the cohomology of incidence alge- bras of partially ordered sets”. In: Communications in Algebra 18.3 (1990), pp. 873–887. [LR20] Fosco Loregian and Emily Riehl. “Categorical notion s of ﬁbration”. In: Expositiones Mathematicae 38.4 (2020), pp. 496–514. [Lur17] Jacob Lurie. Higher Algebra

work page 1990

[3] [12]

2024, e24

Cambridge University Press. 2024, e24. [Huy06] Daniel Huybrechts. Fourier-Mukai transforms in algebraic geometry. Claren- don Press,

work page 2024

[4] [183]

Posets, regular CW complexes and B ruhat order

[Bjö84] Anders Björner. “Posets, regular CW complexes and B ruhat order.” In: Eur. J. Comb. 5.1 (1984), pp. 7–16. [BK05] Eric Babson and Dmitry N Kozlov. “Group actions on pos ets”. In: Journal of Algebra 285.2 (2005), pp. 439–450. [Bon06] Alexey Bondal. Derived categories of toric varieties, Convex and algebrai c geometry, Oberwolfach conference reports, vol. 3

work page 1984

[5] [1994]

Lectures on noncommutative resol utions

[Wem12] Michael Wemyss. “Lectures on noncommutative resol utions”. In: arXiv preprint arXiv:1210.2564 (2012). [Woo98] David Woodcock. “Cohen–Macaulay complexes and Kos zul rings”. In: Journal of the London Mathematical Society 57.2 (1998), pp. 398–410. [Yan05] Kohji Yanagawa. “Dualizing complex of the incidenc e algebra of a ﬁnite regular cell complex”. I...

arXiv 2012

[6] [2006]

Helices, representations of quiv ers and Koszul algebras

[Bon90] Alexei I Bondal. “Helices, representations of quiv ers and Koszul algebras”. In: Helices and vector bundles 148 (1990), pp. 75–95. [BWW05] Anders Björner, Michelle Wachs, and Volkmar Welker . “Poset ﬁber the- orems”. In: Transactions of the American Mathematical Society 357.5 (2005), pp. 1877–1899. [CLS24] David A Cox, John B Little, and Henry K S...

work page 1990

[7] [2017]

Rings with several objects

[Mit72] Barry Mitchell. “Rings with several objects”. In: Advances in Mathematics 8.1 (1972), pp. 1–161. [Pol95] Patrick Polo. “On Cohen-Macaulay posets, Koszul al gebras and certain modules associated to Schubert varieties”. In: Bulletin of the London Mathematical Society 27.5 (1995), pp. 425–434. [Pol97] Alexander Polishchuk. “Perverse sheaves on a tria...

arXiv 1972

[8] [2024]

Moduli of McKay quiver representations I: the coherent component

[CMT07a] Alastair Craw, Diane Maclagan, and Rekha R Thomas. “Moduli of McKay quiver representations I: the coherent component”. In: Proceedings of the London Mathematical Society 95.1 (2007), pp. 179–198. [CMT07b] Alastair Craw, Diane Maclagan, and Rekha R Thomas. “Moduli of McKay quiver representations II: Gröbner basis techniques”. In: Journal of Alge- ...

arXiv 2007