{"id":"4b544f7a-21b2-4ebe-b559-abb61dceef23","arxiv_id":"2411.15359","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Bounded t-deformations of a bounded t-dg-category are equivalent to dg-deformations of its category of derived injectives, and HH^n classifies them for n at least 2.","lead":"The authors prove that two deformation problems in higher category theory, deforming structured triangulated categories and deforming their derived-injective subcategories, are equivalent, completing a program started in earlier papers. A consequence is that higher Hochschild cohomology classes count these deformations, giving a concrete handle on formal deformations in derived geometry and noncommutative algebra.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The main deformation equivalence is only proved for square-zero kernels; the reduction to higher nilpotency is asserted via [LV06] and not demonstrated in the t-dg setting.","rationale":"I read the paper in good faith and followed the proof of the main theorem. The central claim is a composition of deformation equivalences: Proposition 2.5.2, Corollary 2.6.18, and Theorem 2.7.4. The weakest point is indeed in Section 2.6, where the entire lifting machinery is built on the base change exact triangle (42). That triangle is only constructed when the kernel K is nilpotent of degree 2, because the proof needs K to carry a dg-S-action. For higher nilpotency the paper gives only a reference to [LV06] for a composition reduction, without proving the analogous statement for t-deformations. This is not an internal contradiction, but it is a genuine gap in the proof of Corollary 2.7.5 as stated for all allowed theta. The Hochschild corollary C.5 remains safe, since its theta has square-zero kernel, so my concern does not overturn the paper's main motivation. The reader's CONDITIONAL verdict is appropriate; I do not see a reason to move it.","tokens_in":64354,"tokens_out":9093,"duration_ms":88089,"concrete_test":"Work out the n=3 case with R=k[epsilon]/(epsilon^3), S=k, theta the quotient, for a concrete bounded t-dg-category such as B=Perf(k). Try to construct the exact triangle (42) directly for K=(epsilon) by checking whether RHom_R(K, Y(A)) lies in the essential image of Y^(S); if it does not, verify whether the claimed composition of two square-zero deformations nevertheless yields the equivalence of Corollary 2.7.5 for this theta. A complete proof or a counterexample for this case settles whether the reduction is valid.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Corollary 2.7.5 and its supporting Theorem 2.6.17 depend on the base change exact triangle (42) and its heart-level consequence (43). Corollary 2.6.3 explicitly uses that K has a dg-S-action, an action that exists only when K^2=0. For a higher-nilpotency kernel, e.g. theta: k[epsilon]/(epsilon^3) -> k with K=(epsilon), the preimage action is not well-defined: the preimage 0 of 0 in k acts on epsilon as epsilon*epsilon = epsilon^2 != 0. The paper handles this by asserting that a deformation along theta is a composition of square-zero deformations, quoting [LV06, Remark 6.2]; no proof is given for t-deformations, and that reference concerns abelian deformations. This is the load-bearing gap: the Conventions allow K nilpotent of arbitrary order n>0, so the main equivalence is proved only for n<=2 as written. The Hochschild corollary C.5 is not affected, since theta_{2-n} has square-zero kernel, but the theorem as stated overreaches.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","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.","tokens_in":64524,"tokens_out":5334,"duration_ms":51099,"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":[{"comment":"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.","section":"§2.6, opening paragraph; Corollary 2.6.3; Corollary 2.7.5"}],"minor_comments":[{"comment":"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.","section":"Definition 1.1.1 and throughout §1.1–§1.3"},{"comment":"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.","section":"Example 2.3.4"},{"comment":"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.","section":"Remark 2.5.1"}],"recommendation":"major_revision","confidential_remarks":"The sole substantive concern is the unproved passage from square-zero kernels to arbitrary nilpotent kernels in the main deformation equivalence. This is fixable either by supplying the missing reduction or by restricting the statements of the main theorems to nilpotency degree at most two; I do not see grounds for rejection. The paper's dependence on the unpublished reference [GL V] is substantial and may deserve editorial attention."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","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).\n\nThe 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.\n\nI 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.\n\nWho 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.","headline":"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.","tokens_in":65139,"tokens_out":3169,"would_cite":true,"duration_ms":30133,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["18G80","16E45","13D10","18G25"],"pacs":[],"model":"deepseek-v4-flash","headline":"Bounded t-deformations of a dg-category are classified by dg-deformations of its derived injectives.","keywords":["t-structures","dg-categories","derived injectives","deformation theory","Hochschild cohomology","homotopy ind-dg-completion","triangulated categories","curvature"],"falsifier":"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 $\\theta_{-1}$.","tokens_in":64101,"feed_emoji":"","tokens_out":7464,"duration_ms":62947,"temperature":0.7,"pith_summary":"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.","feed_headline":"t-deformations match dg-deformations of derived injectives","feed_subtitle":"The match gives higher Hochschild cohomology a direct deformation-theoretic meaning, with no curvature obstruction.","key_machinery":"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.","core_discovery":"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.","pith_inferences":["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."],"forward_implications":["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."],"supporting_citations":[{"why":"Supplies the reconstruction equivalence between derived-injective dg-categories and t-dg-categories that the deformation equivalences extend.","marker":"[GL V21]"},{"why":"Establishes the base change functors, t-deformations, dg-deformations, and the fact that dg-deformations of derived injectives induce t-deformations.","marker":"[GL V]"},{"why":"Provides the abelian deformation-theoretic model and the reduction of higher nilpotency degree to square-zero deformations used in Section 2.6.","marker":"[LV06]"},{"why":"Compares the Hochschild complex of an abelian category with that of its bounded derived dg-category, the special case behind Theorem C.1.","marker":"[LV05]"},{"why":"Gives the Morita-type criterion for isomorphisms of Hochschild complexes used to prove Theorem C.4.","marker":"[Kel03]"},{"why":"Supplies the A-infinity deformation argument adapted in Lemma C.2, identifying Hochschild cocycles with dg-deformations along $\\theta_{2-n}$.","marker":"[Low08]"}],"fun_headline_variants":["t-deformations equal dg-deformations of derived injectives","No curvature: t-deformations from derived injectives","Hochschild cohomology governs t-deformation theory","Deformations of t-categories: an injective bridge","Derived injectives unlock t-deformation theory"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"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.","fun_headline_variants_meta":{"raw":{"variants":["t-deformations equal dg-deformations of derived injectives","No curvature: t-deformations from derived injectives","Hochschild cohomology governs t-deformation theory","Deformations of t-categories: an injective bridge","Derived injectives unlock t-deformation theory"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001228,"raw_usage":{"total_tokens":5043,"prompt_tokens":941,"completion_tokens":4102,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":557,"completion_tokens_details":{"reasoning_tokens":4023}},"tokens_in":557,"tokens_out":4102,"duration_ms":26809,"temperature":1.0,"reasoning_tokens":4023,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:23:24.457206+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"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 $\\theta_{-1}$.","supporting_citations":[],"review_version":1}