REVIEW 4 major objections 4 minor 7 references
Dualizing complexes and $t$-structures for algebraic stacks
T0 review · 4 major / 4 minor · reviewed 2026-08-02 · deepseek-v4-flash
Pith's one-line read For a separated tame Deligne–Mumford stack of finite presentation over a field, a dualizing complex always exists.
desk verdict Novel existence theorem for dualizing complexes on tame DM stacks, but the abstract overclaims a t-structure classification absent from the body, and the proof leans on unpublished compactification results. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the upper-shriek functor f^! on the 2-category S_e, constructed from the right adjoint f^× of Rf_* by Neeman's formalism for concentrated morphisms, together with the identification f^× ≅ f^! for universally quasi-proper morphisms. The argument uses Nagata compactification to factor f as a dominant flat monomorphism followed by a universally quasi-proper finite-type morphism, then uses base-change isomorphisms and the reduction to algebraic spaces. A dualizing complex is defined smooth-locally on the lisse-étale site, so the proof checks the property on étale schemes covering the stack and descends.
What would settle it
Find a separated tame Deligne–Mumford stack of finite presentation over a field that provably has no dualizing complex; Corollary 1.2 would fail. Short of that, exhibiting a morphism in S_e that lacks the required Nagata compactification would undercut the construction of f^!.
Extended reading notes
Core claim
The central claim, Theorem 4.8, is that the operation K ↦ f^!K preserves dualizing complexes for Deligne–Mumford morphisms of finite presentation within the 2-category S_e of Noetherian algebraic stacks with quasi-affine diagonals that have Nagata compactifications. Theorem 1.1 is the special case where S_e consists of tame Noetherian Deligne–Mumford k-stacks with separated diagonal. From it, Corollary 1.2 concludes that every separated tame Deligne–Mumford k-stack of finite presentation over a field admits a dualizing complex; in characteristic zero this covers every separated Deligne–Mumford stack of finite presentation. A relative variant (Corollary 1.3) states that a tame proper Deligne–
Load-bearing premise
The proof depends on a Nagata compactification theorem for tame Deligne–Mumford stacks quoted from a forthcoming paper; if that theorem is not available in the stated generality, the factorization that defines f^! and drives the reduction is missing.
Editorial extensions
If this is right
- Every separated tame Deligne–Mumford stack of finite presentation over a field carries a dualizing complex, even when not proper.
- In characteristic zero, every separated Deligne–Mumford stack of finite presentation over a field carries a dualizing complex.
- For tame proper Deligne–Mumford morphisms, the right adjoint f^× of derived pushforward preserves dualizing complexes (Corollary 1.3).
- Because these complexes are pseudocoherent with bounded cohomology, they can serve as input for duality, residue, and singularity-theoretic tools on stacks.
- The existence result removes properness restrictions that limited earlier approaches.
Reading between the lines
- The abstract advertises a classification of all tensor t-structures on D^b_coh, but the body contains no proof of that classification; this extraction only treats the dualizing-complex existence results, and the classification should be regarded as unproved here.
- If the quoted forthcoming Nagata compactification theorem for tame Deligne–Mumford stacks is not available in the stated generality, the main reduction collapses; a natural test is to find an independent proof of that compactification or a counterexample.
- The same strategy could apply to other classes of stacks once Nagata compactifications and étale descent for dualizing complexes are known; tameness is used for linearly reductive stabilizers, so non-tame positive-characteristic stacks are a natural boundary.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper develops a notion of dualizing complex on algebraic stacks via the lisse-étale site (Definition 3.4) and proves, under a 2-categorical framework S_e admitting Nagata compactifications (Notation 4.1), that for a Deligne–Mumford morphism f: Y → X in S_e of finite presentation, f^! preserves dualizing complexes (Theorem 4.8). The main special case is Theorem 1.1: for a separated finite-type morphism between tame Deligne–Mumford k-stacks with X Noetherian and separated diagonal, f^! of a dualizing complex is dualizing. Corollary 1.2 concludes existence for separated tame DM k-stacks of finite presentation, and Corollary 1.3 gives a proper version using f^×. The proof reduces to algebraic spaces via étale/smooth descent and uses base-change isomorphisms from [Nee23, Theorem 1.8] and the algebraic-space result [LM22, Lemma 2.7]. The abstract also promises a classification of all tensor t-structures on D^b_coh, but no such classification appears in the body.
Significance. If the existence results are correct, they constitute a significant advance: dualizing complexes are obtained for separated tame Deligne–Mumford stacks without properness constraints, a natural and useful extension of the scheme/algebraic-space theory. The strategy of reducing to algebraic spaces through Neeman's f^! formalism is elegant and potentially influential. The paper is clearly written and carefully distinguishes the f^! and f^× functors. However, the proof rests on two substantial external inputs—[Ryd26, Theorem B] (a forthcoming compactification theorem) and [LM22, Lemma 2.7] (an algebraic-space duality preservation result)—neither of which is stated or proved. The main proof also contains a terse truncation argument in Lemma 4.6 that appears insufficient as written. These issues are fixable but currently prevent full verification.
major comments (4)
- [Lemma 4.3 and Proposition 4.7] The construction of the 2-category S_e in Lemma 4.3 and the compactification reduction in Proposition 4.7 depend entirely on [Ryd26, Theorem B], a forthcoming and not yet published result. In particular, Lemma 4.3 asserts that S_e of tame Noetherian DM k-stacks satisfies Notation 4.1, and the proof says only that Nagata compactifications exist by [Ryd26, Theorem B] and that tameness coincides with 'strictly tameness' in the equicharacteristic setting. None of these hypotheses or statements is spelled out. If [Ryd26, Theorem B] has additional hidden hypotheses (e.g., characteristic zero, quasi-projective diagonal, or only for strictly tame stacks), then Lemma 4.3 is false and the factorization f = g∘j used in Proposition 4.7 collapses. Since this step is the bridge from the quasi-proper case to the general case, Theorem 4.8 and Corollary 1.2 are not verifiable from the material supplied.
- [Lemma 4.5 and Proposition 4.7] The algebraic-space input [LM22, Lemma 2.7] is used in a load-bearing way: in Lemma 4.5 it is the step that upgrades (f')^! of a dualizing complex on U to a dualizing complex on Y×_X U, and in Proposition 4.7 it is used again for the étale fiber products. The lemma is not stated, and its hypotheses are not explained. Without a statement or proof, the reduction to the algebraic-space case is not self-contained. This is especially important because the entire proof strategy is to reduce to algebraic spaces; the reader must be able to check that the cited lemma applies to the morphisms in question.
- [Lemma 4.6] The proof of Lemma 4.6 does not convincingly establish that f^× preserves D^+_qc. The argument assumes 'L ∈ D^+_qc(Y) satisfying H^j(L)=0 if j≥c', i.e., that L is bounded above, which is not the condition for membership in D^+_qc; the final conclusion H^i(f^×L)=0 for i≤c−B does not follow from the displayed Hom-vanishing in the way stated. Since Lemma 4.6 is used in Proposition 4.7 to conclude f^×K ∈ D^+_qc(X) before identifying f^×K with f^!K, this gap affects a step in the main proof. The statement is a standard fact (cf. [Sta26, Tag 0E56]), but the proof needs to be rewritten correctly or replaced by a precise citation.
- [Abstract] The abstract states as an application that the paper 'classifies all tensor t-structures on their bounded derived category of coherent sheaves'. No such classification, or even a statement of a theorem about t-structures, appears anywhere in the body of the manuscript. This is a serious overclaim. The authors should either add the promised classification theorem (with proof) or remove the sentence from the abstract. Since the existence of dualizing complexes does not by itself classify all tensor t-structures, this is not a harmless omission.
minor comments (4)
- [Proposition 4.7] Typo: 'of presentation' should be 'of finite presentation'. Also, the phrase 'Deligne-Mumford morphism (f:X→S)∈S_e' uses X for the source while the rest of the paper uses Y; this is confusing and should be standardized.
- [Example 4.2 and Lemma 4.3] The references to [Ryd26, Theorem B] and [Ryd26, Theorem F] are both used, but the numbering and content of these theorems are not described. Since one is a key dependency, please give a precise statement or at least a summary of the hypotheses and conclusions.
- [Remark 4.4] This remark compiles statements from [Nee23, Theorem 1.8] but does not indicate which parts require D^+_qc and which hold on all D_qc. Adding that information would clarify the use of Lemma 4.6.
- [Section 3, Definition 3.4] The definition of dualizing complex on a stack is given 'smooth locally', but the paper does not discuss how this behaves under arbitrary base change or how it depends on the choice of smooth cover. Proposition 3.8 addresses the cover independence, but a brief discussion would improve readability.
Circularity Check
No significant circularity; the main theorem is conditional on external compactification and duality results, not on its own conclusions.
full rationale
The derivation is input-independent. Theorem 4.8 / Theorem 1.1 are proved from the upper-shriek base-change formalism [Nee23, Theorem 1.8], the algebraic-space duality fact [LM22, Lemma 2.7], and Nagata compactification of tame DM stacks quoted as [Ryd26, Theorem B]. None of these is the target theorem, and none is derived from the paper's conclusions. The dualizing-complex definition (Definition 3.4) is smooth-local, and Proposition 3.8 / Corollary 3.9 are standard descent facts, not restatements of the existence theorem. The only overlap with the author's own prior work is [HLLP25, Lemma 2.3] cited in §2.4 for concentration of morphisms between tame DM stacks; this is redundant with [HR15, Theorem 2.1] already quoted there, so it is a minor and non-load-bearing self-citation. No parameter is fitted and no 'prediction' is definitionally equal to an input. The abstract's advertised classification of tensor t-structures does not appear in the body, and the central theorem is conditional on the unpublished [Ryd26, Theorem B]; both are concerns about completeness and verification, not circularity.
Assumptions & free parameters
assumptions (6)
- domain assumption Nagata compactification for separated finite-type morphisms of tame Noetherian DM k-stacks with separated diagonal ([Ryd26, Theorem B]).
- domain assumption Neeman's f^! functor and its base-change/composition identities for the 2-category S_e ([Nee23, Notation 1.1, Theorem 1.8]).
- domain assumption [LM22, Lemma 2.7]: for morphisms of Noetherian algebraic spaces relevant here, f^! of a dualizing complex is dualizing.
- domain assumption Tameness coincides with 'strict tameness' in the equicharacteristic setting, allowing [Ryd26] to apply.
- domain assumption Morphisms between tame DM stacks are concentrated ([HLLP25, Lemma 2.3]).
- standard math For separated étale morphisms between Noetherian schemes, pullback and upper shriek agree on D+_qc ([Sta26, Tag 0FWI]).
Cite this review
Pith. "Pith review of Dualizing complexes and $t$-structures for algebraic stacks." pith.science (2026). https://pith.science/paper/Z6TD5ZR7
@misc{pith2026260220742,
author = {Pith},
title = {Pith review of: Dualizing complexes and $t$-structures for algebraic stacks},
year = {2026},
howpublished = {\url{https://pith.science/paper/Z6TD5ZR7}},
note = {Machine review of arXiv:2602.20742}
}
abstract
This work is concerned with dualizing complexes on algebraic stacks. We show their existence for suitable Deligne--Mumford stacks. As an application, we classify all tensor $t$-structures on their bounded derived category of coherent sheaves.
Reference graph
Works this paper leans on
-
[25]
Grothendieck duality for Deligne-Mumford stacks
[Nir09] Fabio Nironi. Grothendieck duality for Deligne-Mumford stacks. arXiv:0811.1955,
arXiv 1955
-
[2005]
Providence, RI: American Mathematical Society (AMS),
Proceedings of the 2005 Summer Research Institute, Seattle, W A, USA, July 25–August 12, 2005, pages 259–271. Providence, RI: American Mathematical Society (AMS),
2005
-
[2009]
Birational geometry of deligne-mumford stacks
[KT23] Andrew Kresch and Yuri Tschinkel. Birational geometry of deligne-mumford stacks. arXiv:2312.14061,
-
[2017]
Compact approximation and descent for algebraic stacks
[HLLP25] Jack Hall, Alicia Lamarche, Pat Lank, and Fei Peng. Compact approximation and descent for algebraic stacks. arXiv:2504.21125,
-
[2023]
[LM22] Shiji Lyu and Takumi Murayama. The relative minimal model program for excellent algebraic spaces and analytic spaces in equal characteristic zero. arXiv:2209.08732,
-
[2025]
Algebraic groups and compact generation of their derived categories of representations.Indiana Univ
DUALIZING COMPLEXES FOR ALGEBRAIC STACKS 11 [HR15] Jack Hall and David Rydh. Algebraic groups and compact generation of their derived categories of representations.Indiana Univ. Math. J., 64(6):1903–1923,
1903
-
[2026]
[Gro10] Philipp Gross.Vector bundles as generators on schemes and stacks
Id/No e70409. [Gro10] Philipp Gross.Vector bundles as generators on schemes and stacks. PhD thesis, Düsseldorf, Univ., Diss., 2010,
2010
Reviewed August 2, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.