Deformations of triangulated categories with t-structures via derived injectives

Francesco Genovese; Julie Symons; Michel Van den Bergh; Wendy Lowen

REVIEW 1 major objections 3 minor 12 references

Deformations of triangulated categories with t-structures via derived injectives

T0 review · 1 major / 3 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Bounded t-deformations of a dg-category are classified by dg-deformations of its derived injectives.

desk verdict The paper completes a real program and the Hochschild corollary holds, but the main theorem is only proved for square-zero kernels despite being stated for arbitrary nilpotent ones. read the letter →

arxiv 2411.15359 v1 pith:XCT4BHP2 submitted 2024-11-22 math.CT math.KT

Francesco Genovese , Wendy Lowen , Julie Symons , Michel Van den Bergh This is my paper · ORCID

classification math.CTmath.KT MSC 18G8016E4513D1018G25

keywords t-structuresdg-categoriesderivedinjectivesdeformationtheoryHochschildcohomologyhomotopyind-dg-completiontriangulatedcategoriescurvature

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 tries to prove that deforming a triangulated category equipped with a t-structure (a compatible way of separating objects into negative and positive halves) is exactly the same problem as deforming a simpler dg-category built from its derived injective objects. The main theorem states that, for an essentially small bounded t-dg-category, bounded t-deformations along a suitable dg-ring morphism correspond naturally in the morphism to dg-deformations of the derived injectives of its homotopy ind-dg-completion. Because that derived-injective dg-category is cohomologically concentrated in nonpositive degrees, its dg-deformations carry no curvature, so the usual curvature obstruction to interpreting Hochschild cohomology as deformation theory disappears. The concrete payoff is Corollary C.5: for every $n \geq 2$, the Hochschild cohomology group $HH^n_{dg}(A)$ is isomorphic to the set of bounded t-deformations of $A$ along $\theta_{2-n}: k[\epsilon]/(\epsilon^2) \to k$ with $|\epsilon| = 2-n$, up to equivalence.

What carries the argument

The load-bearing object is the homotopy ind-dg-completion $Ind_{dg,Q,+}(A)$, built from filtered homotopy dg-colimits of representable objects; it is a dg-enhancement of the derived category $D(A)$ and plays the role that the ind-completion plays in abelian deformation theory. The paper shows that the t-structure on $A$ extends to this completion with a t-exact Yoneda embedding, and that the completion is a left bounded locally coherent Grothendieck t-dg-category. Its full dg-subcategory $DG-Inj$ of derived injectives is cohomologically concentrated in nonpositive degrees, which is what removes curvature. The second workhorse is the base change exact triangle (42), obtained from the exact triangle $K \to R \to S$ with $K$ the square-zero kernel of $\theta$; the triangle lets the authors transfer coproducts, products, and compactness properties from the deformed category back to its heart. The whole argument proceeds by proving equivalences of deformation pseudofunctors stepwise, then combining them in Corollary 2.7.5.

What would settle it

Construct a t-deformation along a $\theta$ whose kernel is nilpotent of degree three and test whether it factors into two square-zero deformations; a failure would invalidate the main equivalence for general $\theta$ while leaving Corollary C.5 intact. Equivalently, compute $HH^3_{dg}(D^b(k))$ and compare it with bounded t-deformations of $D^b(k)$ along $\theta_{-1}$.

Watch

Extended reading notes

Core claim

The central claim is a natural-in-$\theta$ equivalence of deformation pseudofunctors, $Def^{t,b}_A(\theta) \cong Def^{dg}_{DG-Inj(Ind_{dg,Q,+}(A))}(\theta)$, for any essentially small strongly pretriangulated $S$-linear dg-category $A$ with a bounded t-structure and a suitable morphism $\theta: R \to S$ of commutative dg-rings. The left-hand side is the problem of lifting the t-dg-category $A$ over $\theta$, while the right-hand side is the problem of deforming the dg-category of derived injectives of its homotopy ind-dg-completion. Since that derived-injective dg-category is cohomologically concentrated in nonpositive degrees, its dg-deformations have zero curvature, and the authors conclude that the Hochschild complex of $A$ governs the deformation theory of $A$. The theorem is assembled from three compatible equivalences: one between dg-deformations of derived injectives and t-deformations with enough derived injectives, one between those and left bounded locally coherent Grothendieck t-deformations, and one between the latter and bounded t-deformations of the original category.

Load-bearing premise

Every deformation with a kernel nilpotent of degree greater than two is assumed to decompose into square-zero deformations, and the paper cites this reduction rather than proving it; if that decomposition fails for some $\theta$, the main equivalence for arbitrary nilpotent kernels is not established, even though Corollary C.5 only needs the square-zero case.

Editorial extensions

If this is right

Bounded t-deformations of a bounded t-dg-category are completely classified by dg-deformations of its derived injectives, naturally in the base change $\theta$.
For every $n \geq 2$, the Hochschild cohomology group $HH^n_{dg}(A)$ counts bounded t-deformations along $\theta_{2-n}$, giving higher Hochschild cohomology a direct deformation-theoretic reading.
Because the derived-injective dg-category is nonpositively graded, the curvature problem that complicates curved $A_\infty$-deformations does not arise in this t-structured setting.
The equivalence of Theorem 1.6.6 says essentially small strongly pretriangulated bounded t-dg-categories are the same data as left bounded locally coherent Grothendieck t-dg-categories, via $h-proj_+$ and $hfp^b$.
In characteristic zero, the shifted Hochschild complex controls t-dg-deformations through Maurer-Cartan elements, as stated in Remark C.6.

Reading between the lines

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

These editorial inferences go beyond the paper: the curvature-free mechanism is likely a feature of t-structures rather than of this particular dg-model, so analogous classifications should hold for any deformation problem whose derived injectives can be made nonpositively graded.
The base change exact triangle suggests an explicit obstruction theory: the tangent space of bounded t-deformations should be $HH^2_{dg}(A)$, with higher obstructions living in $HH^{n+1}_{dg}(A)$, and one could try to write down the resulting Maurer-Cartan equation explicitly.
A natural stress test is to verify that the composed square-zero deformations used for higher nilpotency degree assemble without hidden signs into the predicted higher Hochschild cohomology groups.

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

Here's my read. The main equivalence in Corollary 2.7.5 is genuinely new: it gives the missing converse direction of the deformation equivalence, and Corollary C.5 gives a deformation-theoretic reading of HH^n for n≥3 that was previously absent. The B∞-quasi-isomorphism between the Hochschild complexes (Theorem C.4) is a solid piece of work, and the construction of the induced t-structure on the dg-derived category in §1.5, plus the equivalence in Theorem 1.6.6, are real contributions. The paper is careful and flags its own limitations, e.g. non-degeneracy failure in Example 1.5.8 and the non-equivalence of the right arrow in (40).

The soft spot is real and located precisely: §2.6 begins by assuming K^2=0, and the main theorem's generality over arbitrary nilpotent kernels rests on the sentence that higher nilpotency reduces to this case by composing deformations, quoting [LV06, Remark 6.2]. That reference proves the reduction for abelian deformations, not for t-deformations. So Corollary 2.7.5 as stated overreaches: as written, the t-deformation equivalence is proved for square-zero base changes only. The Hochschild corollary is unaffected because θ_{2-n} has square-zero kernel, so the headline application stands. A second practical issue: several load-bearing reconstruction results are from the unpublished [GLV]. That is a footgun for referees and readers, though not a mathematical defect if those results are correct.

I don't see circularity or fitting-to-data. The proofs are detailed and the text takes care to distinguish what is known from what is new.

Who is this for? People working in dg-category deformation theory and noncommutative geometry, especially the deformation-theoretic meaning of higher Hochschild cohomology. The paper deserves a serious referee: it should go to peer review, not be desk-rejected. But the referee report should require the authors to either restrict the main theorem (and the diagram in Cor 2.7.5) to square-zero kernels and leave higher nilpotency as a separate claim with proof, or actually prove the reduction. It would also help to restate the needed results from [GLV] or otherwise make them verifiable.

Referee Report

1 major / 3 minor

Summary. This paper completes a series on deformations of pretriangulated dg-categories with t-structures. It extends a t-structure on an essentially small strongly pretriangulated dg-category A to its dg-derived category Inddg,Q,+(A) (Theorem 1.5.3), identifies the resulting class of left bounded locally coherent Grothendieck t-dg-categories (Theorem 1.6.6), and then proves several equivalences of deformation pseudofunctors. The central result, Corollary 2.7.5, identifies bounded t-deformations Def^{t,b}_B with dg-deformations of derived injectives Def^{dg}_{DG-Inj(h-proj+(B))}. As a consequence, Corollary C.5 interprets HH^n_dg(A) for n≥2 as bounded t-deformations along θ_{2-n}: k[ε]/(ε^2)→k. A Hochschild-level B∞-quasi-isomorphism C(A)≅C(DG-Inj(Inddg,Q,+(A))) is proved in Theorem C.4. The proof of the main deformation equivalence for general nilpotent kernels is not complete in the text: §2.6 assumes K^2=0, and the reduction to higher nilpotency is delegated to [LV06, Remark 6.2] without proof in the t-dg setting.

Significance. The paper is a substantial contribution if the main equivalence holds. It supplies the missing converse to earlier work [GL V21] and [GL V], identifies the correct big t-dg-category, and avoids the curvature obstruction by landing in nonpositively graded derived injectives. The proofs are detailed, and the authors are explicit about limitations: non-degeneracy is not preserved in Example 1.5.8, and the right-hand arrow in diagram (40) is not an equivalence in general. I found no circularity: Corollary 2.7.5 is not assumed as input, and Corollary C.5 uses the independent B∞-quasi-isomorphism of Theorem C.4. However, because the stated generality of Corollary 2.7.5 is not established, the significance is conditional on fixing the square-zero-to-higher-nilpotency reduction or restricting the statement.

major comments (1)

[§2.6, opening paragraph; Corollary 2.6.3; Corollary 2.7.5] The main deformation equivalence is proved only for kernels K with K^2=0. Corollary 2.6.3 and the base change exact triangle (42) require K to carry a dg-S-action, and the text defines this action through preimages under θ; for K=(ε) in θ: k[ε]/(ε^3)→k, the preimage action is not well-defined because 0·ε = ε^2≠0. The Conventions allow nilpotency of arbitrary order n>0, and Corollary 2.7.5 is stated for all θ satisfying those conventions. The reduction to square-zero kernels by composing deformations is asserted with a citation to [LV06, Remark 6.2], which concerns abelian deformations; no transfer argument to t-deformations is provided, and Proposition A.4 repeats the same assumption. This is load-bearing: as written, Theorem 2.6.17, Corollary 2.6.18, Theorem 2.7.4 and Corollary 2.7.5 hold only for n≤2. Corollary C.5 is unaffected, since its θ_{2-n} has square-zero kernel, but the main theorem overreaches.

minor comments (3)

[Definition 1.1.1 and throughout §1.1–§1.3] The typesetting of weighted (co)limit symbols is corrupted in many displayed formulas (for example 'lim←/leftr⫯g⊸tl⫯ne'), which makes the definitions substantially harder to read.
[Example 2.3.4] The text cites Example 1.4.9 as justification that compactly generated dg-categories have enough derived injectives, but Example 1.4.9 concerns a degenerate t-structure on D(k[u,u^{-1}]); the reference appears to be a cross-reference error.
[Remark 2.5.1] The notation Def_B(R) is used for the W-groupoid after the introduction of additional universes V and W, which may be confused with the pseudofunctor Def_B defined earlier; please disambiguate the notation.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the main deformation equivalence is a new composition of previously established independent results; the only flagged issue is an asserted reduction for higher-nilpotency kernels, which is a generality gap rather than a circular step.

full rationale

The paper's central claims are not circular. The target equivalence Def^{t,b}_B ≅ Def^{dg}_{DG-Inj(h-proj+(B))} (Corollary 2.7.5) is assembled from three separate pseudofunctor equivalences (Proposition 2.5.2, Corollary 2.6.18, Theorem 2.7.4), each proved from the definitions of t-deformations and dg-deformations and from prior reconstruction results ([GLV21], [GLV]). These citations are load-bearing, but they are independent prior theorems about t-structures and twisted complexes, not restatements of the present target; the paper adds the missing equivalence between bounded t-deformations and dg-deformations of derived injectives. The Hochschild corollary C.5 is a genuine consequence: Theorem C.4 establishes a B-infinity quasi-isomorphism C(A) ≅ C(DG-Inj(Inddg,Q,+(A))) via Keller's Morita-invariance criterion, and Lemma C.2 is the standard cocycle-to-A-infinity-deformation bijection, with no curvature because the derived-injective dg-category is cohomologically concentrated in nonpositive degree. No fitted parameter is renamed as a prediction, and no uniqueness theorem is imported from the authors to force a choice. The flagged limitation in Section 2.6 — the reduction from arbitrary nilpotent kernel to square-zero kernel is asserted by citation to [LV06, Remark 6.2] rather than proved in the t-dg setting — is a correctness or overreach concern about the statement's full generality, not a circular step: the square-zero case is proved independently and is sufficient for Corollary C.5. Hence there is no significant circularity.

Assumptions & free parameters 0 free parameters · 5 assumptions · 2 invented entities

The main claims rest on the paper's standing conventions for theta, on the universe axioms, on the class of strongly pretriangulated bounded t-dg-categories, and on several previously established results by the authors. There are no fitted constants; the free parameters list is empty.

assumptions (5)

standard math ZFCU, Zermelo-Fraenkel with Choice and the Universe axiom, with fixed universes U in V in W as needed in Appendix A.
Stated in Conventions and Remark 2.5.1; size conditions are necessary for the homotopy ind-dg-completion properties and for defining deformation pseudofunctors.
domain assumption Standing Conventions on theta: R to S is strictly surjective, concentrated in nonpositive degrees, homotopically coherent, with kernel K nilpotent and S and K finitely presented in Z^0(dgm(R)).
Defines the class of dg-ring morphisms over which deformations are considered; used throughout the paper, including the base change formula and finiteness arguments.
domain assumption In section 2.6, K is nilpotent of degree 2, and higher nilpotency degree is asserted to reduce to this case by composing deformations.
The base change exact triangle (42) requires K to carry a dg-S-action, which the authors state exists when K has square zero. The reduction for general n is quoted from the abelian story in [LV06] rather than proved here.
domain assumption The main theorem assumes B is an essentially small strongly pretriangulated S-linear dg-category with a bounded t-structure, with h-flatness obtained by cofibrant replacement where needed.
This is the object class on the small side of the deformation equivalence; strong pretriangulatedness is needed for Theorem 1.3.7 and is explicitly noted as unclear for merely pretriangulated categories in Remark 1.3.8.
standard math Brown representability in the left bounded setup of Appendix B: T^{>=0} is closed under countable coproducts, T is generated by a set of compact objects, and the cohomological functor H maps coproducts to products and satisfies H composed with tau^{>=0} congruent to H.
Used in Corollary B.2 and Corollary 1.6.5 to prove that left bounded locally coherent Grothendieck t-dg-categories have enough derived injectives.

invented entities (2)

Homotopy ind-dg-completion Inddg,Q(A) independent evidence
purpose: A dg-enhancement of the derived category D(A) built from filtered homotopy dg-colimits of representables; serves as the big category in the deformation equivalence.
It is shown in Theorem 1.3.7 to be quasi-equivalent to the independently existing h-proj(A), so the construction has a checkable identification.
Left bounded locally coherent Grothendieck t-dg-category
purpose: A class of large t-dg-categories serving as the codomain of the equivalence in Theorem 1.6.6 and as the deformation target in section 2.6.
Newly introduced in section 1.6; its utility is proven internally via Corollary 1.6.5 and Theorem 1.6.6, but there is no external falsifiable handle.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Deformations of triangulated categories with t-structures via derived injectives." pith.science (2026). https://pith.science/paper/XCT4BHP2

@misc{pith2026241115359,
  author       = {Pith},
  title        = {Pith review of: Deformations of triangulated categories with t-structures via derived injectives},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XCT4BHP2}},
  note         = {Machine review of arXiv:2411.15359}
}

read the original abstract

This paper provides the final ingredient in the development of the deformation theory of pretriangulated dg-categories endowed with a nice t-structure, which was initiated by the authors and is modeled after the previously developed deformation theory of abelian categories. We show how to extend a t-structure on a pretriangulated dg-category to its dg-derived category so that the Yoneda embedding becomes t-exact. We construct several equivalences between deformation problems; in particular, we prove a deformation equivalence between the bounded t-deformations of a bounded t-dg-category on the one hand, and dg-deformations of the dg-category of derived injective ind-dg-objects on the other hand. Since this latter dg-category is cohomologically concentrated in nonpositive degrees, we do not encounter curvature.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

12 extracted references · 12 canonical work pages

[4]

Formal moduli problems and for- mal derived stacks,

[CG21] D. Calaque and J. Grivaux, “Formal moduli problems and for- mal derived stacks,” in Derived algebraic geometry, ser. Panor. Synth` eses, vol. 55, Soc. Math. France, Paris, 2021, pp. 85–145. [COS19] A. Canonaco, M. Ornaghi, and P. Stellari, “Localizations of the category of A∞ categories and internal Homs,”Doc. Math., vol. 24, pp. 2463–2492,

work page 2021
[6]

Available: https : / / webusers

[Online]. Available: https : / / webusers . imj - prg.fr/~bernhard.keller/publ/dih.pdf. [Kel82] G. M. Kelly, Basic concepts of enriched category theory (London Mathematical Society Lecture Note Series). Cambridge University Press, Cambridge-New York, 1982, vol. 64, p

work page 1982
[8]

Filtered derived categories of curved deformations

arXiv: 2402.08660 [math.KT]. [Low05] W. Lowen, “Obstruction theory for objects in abelian and derived categories,” Comm. Algebra, vol. 33, no. 9, pp. 3195–3223,

work page Pith review arXiv
[12]

Index Hqe(R), 2 cofibrant replacement Q(A), 3 HqeDG-Inj (Π) , 40 Hqet+ (Π), 39 dgm(A), 3 Ho(B∞), 69 t-dg-category, 24 left bounded locally coherent Grothendieck, 33 truncation quasi-functors τ≤n, τ≥n, 25 t-structure (T≤0, T≥0), 19 i-th cohomology H i t , 19 t-exact, left/right, 20 (co)aisle T ≤0 and T ≥0, 21 (non-)degenerate, 19 bounded, left/right, 20 cl...

work page 2020
[1981]

Ast´ erisque, vol

, ser. Ast´ erisque, vol. 100, Soc. Math. France, Paris, 1982, pp. 5–

work page 1982
[2003]

Faisceaux pervers,

[AR94] J. Ad´ amek and J. Rosick´ y,Locally presentable and accessible cate- gories (London Mathematical Society Lecture Note Series). Cam- bridge University Press, Cambridge, 1994, vol. 189, pp. xiv+316. [BBD82] A. A. Beilinson, J. Bernstein, and P. Deligne, “Faisceaux pervers,” in Analysis and topology on singular spaces, I (Luminy,

work page 1994
[2011]

Non-left-complete derived categories

arXiv: 1103.5539 [math.CT]. [Nee18] A. Neeman, The category [T c] op as functors on T b c ,

work page Pith review arXiv
[2015]

The curvature problem for formal and infinitesimal deformations

arXiv: 1505.03698 [math.KT]. [Nee01] A. Neeman, Triangulated categories (Annals of Mathematics Stud- ies). Princeton University Press, Princeton, NJ, 2001, vol. 148, pp. viii+449. [Nee11] A. Neeman, Non-left-complete derived categories ,

work page Pith review arXiv 2001

Show all 12 references

[2018]

The Grothendieck duality theorem via Bousfield’s techniques and Brown representability,

arXiv: 1806.05777 [math.CT]. [Nee96] A. Neeman, “The Grothendieck duality theorem via Bousfield’s techniques and Brown representability,” J. Amer. Math. Soc. , vol. 9, no. 1, pp. 205–236,

arXiv
[2019]

Construction of t-structures and equivalences of derived cate- gories,

[AGV72] M. Artin, A. Grothendieck, and J. -L. Verdier, Th´ eorie des topos et cohomologie ´ etale des sch´ emas. Tome 1: Th´ eorie des topos (Lecture Notes in Mathematics, Vol. 269). Springer-Verlag, Berlin- New York, 1972, pp. xix+525, S´ eminaire de G´ eom´ etrie Alg´ ebriqu...

1972
[2021]

A derived gabriel–popescu theorem for t-structures via derived injectives,

[GR22] F. Genovese and J. Ramos Gonz´ alez, “A derived gabriel–popescu theorem for t-structures via derived injectives,” International Mathematics Research Notices, vol. 2023, no. 6, pp. 4695–4760, Jan. 2022, issn: 1687-0247. [Kel03] B. Keller, Derived invariance of higher str...

2023
[2024]

[LL24] A

arXiv: 2406.04945 [math.KT]. [LL24] A. Lehmann and W. Lowen, Filtered derived categories of curved deformations,

arXiv

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [4]

Formal moduli problems and for- mal derived stacks,

[CG21] D. Calaque and J. Grivaux, “Formal moduli problems and for- mal derived stacks,” in Derived algebraic geometry, ser. Panor. Synth` eses, vol. 55, Soc. Math. France, Paris, 2021, pp. 85–145. [COS19] A. Canonaco, M. Ornaghi, and P. Stellari, “Localizations of the category of A∞ categories and internal Homs,”Doc. Math., vol. 24, pp. 2463–2492,

work page 2021

[2] [6]

Available: https : / / webusers

[Online]. Available: https : / / webusers . imj - prg.fr/~bernhard.keller/publ/dih.pdf. [Kel82] G. M. Kelly, Basic concepts of enriched category theory (London Mathematical Society Lecture Note Series). Cambridge University Press, Cambridge-New York, 1982, vol. 64, p

work page 1982

[3] [8]

Filtered derived categories of curved deformations

arXiv: 2402.08660 [math.KT]. [Low05] W. Lowen, “Obstruction theory for objects in abelian and derived categories,” Comm. Algebra, vol. 33, no. 9, pp. 3195–3223,

work page Pith review arXiv

[4] [12]

Index Hqe(R), 2 cofibrant replacement Q(A), 3 HqeDG-Inj (Π) , 40 Hqet+ (Π), 39 dgm(A), 3 Ho(B∞), 69 t-dg-category, 24 left bounded locally coherent Grothendieck, 33 truncation quasi-functors τ≤n, τ≥n, 25 t-structure (T≤0, T≥0), 19 i-th cohomology H i t , 19 t-exact, left/right, 20 (co)aisle T ≤0 and T ≥0, 21 (non-)degenerate, 19 bounded, left/right, 20 cl...

work page 2020

[5] [1981]

Ast´ erisque, vol

, ser. Ast´ erisque, vol. 100, Soc. Math. France, Paris, 1982, pp. 5–

work page 1982

[6] [2003]

Faisceaux pervers,

[AR94] J. Ad´ amek and J. Rosick´ y,Locally presentable and accessible cate- gories (London Mathematical Society Lecture Note Series). Cam- bridge University Press, Cambridge, 1994, vol. 189, pp. xiv+316. [BBD82] A. A. Beilinson, J. Bernstein, and P. Deligne, “Faisceaux pervers,” in Analysis and topology on singular spaces, I (Luminy,

work page 1994

[7] [2011]

Non-left-complete derived categories

arXiv: 1103.5539 [math.CT]. [Nee18] A. Neeman, The category [T c] op as functors on T b c ,

work page Pith review arXiv

[8] [2015]

The curvature problem for formal and infinitesimal deformations

arXiv: 1505.03698 [math.KT]. [Nee01] A. Neeman, Triangulated categories (Annals of Mathematics Stud- ies). Princeton University Press, Princeton, NJ, 2001, vol. 148, pp. viii+449. [Nee11] A. Neeman, Non-left-complete derived categories ,

work page Pith review arXiv 2001