Enriched categorical aspects of homological perturbation theory

Luk\'a\v{s} Vok\v{r}\'inek

REVIEW 3 major objections 7 minor 7 references

Enriched categorical aspects of homological perturbation theory

T0 review · 3 major / 7 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read The perturbation lemma becomes a functor in any dg-category

desk verdict A genuinely new categorical framing of the perturbation lemma with a real gap in Theorem 11 — the unconstructed 'free closure under perturbations' Q — but Theorems 10, 12–14 stand on explicit arguments. read the letter →

arxiv 2412.21182 v1 pith:SIOIPBWX submitted 2024-12-30 math.CT

Lukáš Vokřínek This is my paper · ORCID

classification math.CT MSC 18G3518D2018N10

keywords homologicalperturbationlemmadg-categoriesstrongdeformationretractionsabsolutelimitsMaurer-Cartanequationenrichedcategorytheorydoublecategoriesfunctoriality

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

The paper seeks to place the homological perturbation lemma—the standard recipe for carrying a perturbation of one chain complex across a homotopy equivalence—inside enriched category theory. It defines perturbations as absolute limits, characterizes strong deformation retractions as vertical arrows in a double category, and proves the lemma by an explicit non-dg-isomorphism $\theta$. The payoff is functoriality: the perturbed strong deformation retraction is produced by a dg-functor, so composing perturbations along composites, iterating the construction, and tensoring transferred structures all yield the same answers. If the claims hold, homological perturbation theory becomes a canonical operation on any dg-category closed under perturbations rather than a case-by-case chain-complex computation.

What carries the argument

The machinery rests on three interlocking objects. The universal dg-category $S$ has two objects $s,t$ and maps $f\colon s\to t$, $g\colon t\to s$, $h\colon s\to s$ of degrees $0,0,1$ subject to $Df=0$, $Dg=0$, $Dh=1-gf$, $1=fg$, $fh=0$, $hg=0$, $hh=0$; a strong deformation retraction in a category $\mathcal{C}$ is exactly a dg-functor $S\to\mathcal{C}$. Perturbations are encoded by adjoining a degree $-1$ generator $\delta$ with $D\delta=-\delta^2$, giving the one-object dg-category $\mathbb{Z}[\delta]$; a $\delta$-perturbation $A_\delta$ is then the $W$-weighted limit (equivalently $W'$-weighted colimit) of the diagram $\mathbb{Z}[\delta]\to\mathcal{C}$, so perturbations are absolute limits. The proof itself is carried out in the localization $S_{\mathrm{ps}}^{\mathrm{loc}}$ at $1+\delta h$, using the two non-dg-isomorphisms $\alpha=1+\delta h$ and $\beta=1+h\delta$: $\alpha$ preserves the $g$-side factorization and $\beta$ preserves the $f$-side, and their combination produces the $\theta$ whose logarithmic derivative factors through $f$ and $g$.

What would settle it

Take a concrete dg-category closed under perturbations, choose a composable pair of strong deformation retractions and a perturbation with $1+\delta h$ invertible, and compare the two strong deformation retractions obtained by (i) composing first and perturbing once and (ii) perturbing each retraction and then composing; a difference at any component would refute Theorem 12 and with it the claimed functoriality. Equally direct: construct the free closure $Q$ for a simple example such as $\mathbf{Ch}$, or prove that no such closure can exist, which would settle whether Theorem 11 is supported.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is Theorem 10: for any strong deformation retraction $(f,g,h)\colon A\to B$ in a dg-category and any perturbation $\delta$ of $A$ with $1+\delta h$ invertible, there exists a non-dg-isomorphism $\theta\colon A\to A_\delta$ whose left logarithmic derivative $D^{\ell}\theta=\theta^{-1}D\theta$ factors through $f$ and $g$. This $\theta$ makes the perturbed diagram fillable and yields the explicit formulas $\hat{f}=f(1+\delta h)^{-1}$, $\hat{g}=(1+h\delta)^{-1}g$, $\hat{h}=h(1+\delta h)^{-1}$, and $\delta'=\hat{f}\delta g$. Theorems 11–14 then assert that this construction is a dg-functor from the localized category of perturbed strong deformation retractions to the category of strong deformation retractions, and that it is compatible with vertical composition of retractions, with iteration of perturbations, and with tensor products. The proof is carried out by writing the source and target of the retraction as the objects $s$ and $t$ of the universal dg-category $S$ and pushing the deformation along the non-dg-isomorphisms $1+\delta h$ and $1+h\delta$.

Load-bearing premise

The load-bearing premise is that there is an object $Q$, the 'free closure under perturbations', with the properties used in the proof of Theorem 11; the paper invokes $Q$ but never constructs it, so if no such object exists the functoriality claim has no proof.

Editorial extensions

If this is right

In any dg-category closed under perturbations, the transfer of a perturbation across a strong deformation retraction is one canonical operation, with the same output regardless of the order in which retractions are composed (Theorem 12).
Iterating the perturbation lemma is associative: perturbing $A$ to $A_\delta$ and then to $A_\varepsilon$ gives the same transferred retraction as perturbing directly to $A_\varepsilon$ (Theorem 13).
The construction respects tensor products: $(F_\ell\otimes F_r)^{\delta_\ell\otimes 1+1\otimes\delta_r}=F_\ell^{\delta_\ell}\otimes F_r^{\delta_r}$, which is exactly the compatibility needed for twisted products and Eilenberg–Zilber-style arguments (Theorem 14).
The perturbed strong deformation retraction is unchanged when the two ingredients $\alpha$ and $\beta$ are applied in the opposite order, so the self-dual formulas for $\hat{f}$, $\hat{g}$, $\hat{h}$, and $\delta'$ give a canonical output even though the intermediate non-dg-isomorphism $\theta$ is not unique.

Reading between the lines

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

The paper leaves the free closure $Q$ of Theorem 11 unconstructed; if such a closure can be built, the same functoriality argument should apply to other universal diagram categories, such as homotopy retracts, yielding a uniform enriched perturbation lemma.
Reading a perturbation as an absolute limit suggests an implementation strategy: compute perturbed complexes as weighted limits, which would make each transfer step a single linear construction rather than a bespoke recursive recipe.
The non-uniqueness of $\theta$ noted in Section 5 means the perturbation lemma is not itself a universal property; comparing different choices of $\theta$ could yield a secondary invariant, possibly detected by the composite $\alpha\beta^{-1}\alpha^{-1}\beta$, which the paper observes is unlikely to be $1$.

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.

Desk Editor's Note

read the letter

Dear Colleague,

Here's my read of Vokřínek's paper. The genuinely new pieces are the absolute limit characterization of perturbations (Prop 4) and the double categorical framework for strong deformation retractions. Theorem 10 is the perturbation lemma with explicit formulas, and the proof is direct and checkable. Theorems 12–14 answer concrete compatibility questions — vertical composition, iteration, tensor products — and are proved by explicit diagram chases, not by appeal to abstract machinery. These parts hold up.

The soft spot is Theorem 11, which claims the perturbation lemma association [S^loc_ps,C] → [S,C] is a dg-functor preserving limits and colimits. The proof invokes "Q, the operator of the free closure under perturbations" and a dg-functor P: S⊗Z[δ] → QS^loc_ps, but Q is never defined, constructed, or given a universal property. The subsequent equivalence [S^loc_ps,C] ≅ [QS^loc_ps,C] for dg-categories closed under perturbations is simply asserted. Without Q, the horizontal functoriality claim is unsupported. The stress-test note is right that Theorems 12–14 do not rely on Q — their proofs are direct — so the central perturbation lemma and the main compatibility results stand. But the paper's stated goal of proving full functoriality is only partially met.

There are minor issues too. Prop 4's weighted-limit construction is sketched rather than fully verified, though the idea is clear. And the proof of Theorem 11 also needs closure conditions on C that the author does not spell out. These are fixable.

The paper is honest: the conjecture about absolute limits is clearly labeled, and there is no circularity or self-citation inflation. The author is a serious thinker, and the gap in Theorem 11 is a missing construction, not a fatal flaw. I would send it to peer review with a request to either construct Q or weaken Theorem 11 to a conditional statement.

For a reading group, I'd say maybe — the direct proofs are instructive, but the paper needs work before it is fully quotable.

Referee Report

3 major / 7 minor

Summary. The paper proposes an enriched-categorical framework for homological perturbation theory. It defines a perturbation δ of an object A in a dg-category by a universal property, identifies perturbations as absolute weighted limits and colimits (Proposition 4), and introduces a double-categorical setting whose vertical arrows are strong deformation retractions (Section 4). The main technical result is Theorem 10, the homological perturbation lemma: for a strong deformation retraction A→B and a perturbation δ of A with 1+δh invertible, explicit formulas give a perturbed strong deformation retraction. The paper then claims functoriality of this construction (Theorem 11) and proves compatibility with vertical composition, iteration, and tensor products (Theorems 12–14).

Significance. If fully established, the framework would give a conceptual, enrichment-agnostic account of homological perturbation theory, and the compatibility theorems are useful for applications such as transferring perturbations along composite homotopy equivalences. The paper gives explicit, checkable formulas in Theorem 10, identifies perturbations as weighted limits, and acknowledges the non-uniqueness of the auxiliary map θ. The direct arguments for Theorems 12–14 are a genuine contribution. However, the central functoriality claim in Theorem 11 rests on an unconstructed 'free closure under perturbations' Q and an asserted equivalence involving Q; without a definition or universal property for Q, that claim is not supported by the manuscript.

major comments (3)

[Section 6, Theorem 11] The proof of Theorem 11 is not complete. It introduces 'Q, the operator of the free closure under perturbations' and a dg-functor P : S⊗Z[δ] → QS^loc_ps, but Q is never defined, constructed, or given a universal property. The argument then asserts an equivalence [S^loc_ps,C] ≅ [QS^loc_ps,C] for dg-categories C closed under perturbations; this equivalence is also asserted without proof. Because the theorem claims a dg-functor between diagram categories, one must specify the action on morphisms, and the non-uniqueness of θ in Theorem 10 makes it nontrivial that a coherent functorial choice exists. Without Q and the equivalence, the association is only defined on objects and the claimed dg-functoriality is unsupported. The authors should either construct Q and prove its universal property, or give an explicit, choice-free definition of the functor and verify naturality directly.
[Section 6, Theorem 11 (colimit claim)] The last sentence of the proof states that the colimit statement follows 'from the same argument but with the perturbation interpreted as a weighted colimit.' This is too terse: preserving limits by P^* and by the weighted limit functor does not automatically imply preservation of colimits by the composite, and the argument would need to identify the appropriate colimit-preserving functors in the colimit version. Since Theorem 11 claims preservation of all limits and colimits, this half of the claim needs a separate justification.
[Section 5, proof of Theorem 10] The proof contains several compressed steps that are load-bearing for the explicit formulas. In particular, the claim that 'Sps(s,s)_1 is generated by the compositions ζ = hδ...δh' and the deduction that the universal solution satisfies ζ = h are stated without derivation. The formulas can likely be verified directly, but as written this part of the proof is more an assertion than a demonstration. I recommend expanding this step so the existence and uniqueness claims for α and β are fully justified.

minor comments (7)

[Abstract] The abstract contains a typo: 'homologic al' should be 'homological'.
[Section 1] In the first paragraph, 'occuring' should be 'occurring', and 'decorated' is used in an informal way; consider 'equipped with a small model structure' or similar.
[Section 3, Proposition 4] The proof transports the weighted-limit construction from Ch to an arbitrary dg-category by representability, but the details of this transport are only sketched. Since this is one of the foundational claims, a few more sentences explaining why the representing object exists and is preserved would improve readability.
[Section 4] The pullback description of vertical morphisms uses the notation c1,c0 in a way that is slightly confusing; clarifying that c1 and c0 are the source and target objects of the vertical morphism would help.
[Theorem 10] The parenthetical '(To be perfectively precise, ...)' contains a typo: 'perfectively' should be 'perfectly'.
[Section 6, Theorem 12 proof] The large diagram in the proof of Theorem 12 is hard to follow; labels such as the middle object and the maps 1−δgh'f2 are not fully explained. Replacing part of the diagram with explicit equations would make the argument much clearer.
[Section 6, after Theorem 11] The sentence 'One may understand the proof of the perturbation lemma in QS^loc_ps, in which case it produces exactly the required dg-functor' is cryptic without a definition of Q; this should be expanded or removed until Q is properly introduced.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the perturbation lemma is proved by explicit construction; the Q-based functoriality proof is a gap, not a circle.

full rationale

The derivation chain is self-contained. Perturbations are first defined by a universal property (Definition 2) and characterized in Proposition 3 via the equation Dr phi = -delta; Proposition 4 then constructs perturbed objects as weighted limits and colimits without assuming the perturbation lemma. The proof of Theorem 10 explicitly constructs alpha = 1 + delta h and beta = 1 + h delta and computes f-hat, g-hat, h-hat and delta-prime from these ingredients, so the conclusion is not used as a premise. Proposition 9 is derived from Proposition 1, whose proof is given in the paper, and Theorems 12-14 are argued directly. The only serious weakness is Theorem 11, which invokes an unconstructed 'free closure under perturbations' Q; that makes the functoriality proof incomplete, but it is a missing existence proof rather than a circular reduction. No fitted parameter is renamed as a prediction, no known result is repackaged as a derivation, and no load-bearing self-citation appears: all cited works are external. Under the stated standard, an honest finding is that the paper exhibits no significant circularity.

Assumptions & free parameters 0 free parameters · 5 assumptions · 1 invented entities

The central claim rests on standard enriched category theory plus one unconstructed ingredient: the free closure under perturbations Q. No numerical free parameters appear. The paper's main non-standard assumption is the unproven existence of Q.

assumptions (5)

standard math The category Ch of chain complexes is a closed symmetric monoidal category and provides enrichment for dg-categories.
Used throughout the paper as the basis for enrichment.
standard math The Yoneda lemma and representable functors behave as usual in enriched categories.
Used in Section 3 to infer invertibility of φ.
domain assumption The cocategory object S (free living strong deformation retraction) exists and [S,C] has the stated vertical maps.
Example 6; this is central to the double categorical framework, but its construction is standard.
ad hoc to paper The 'free closure under perturbations' operator Q exists and behaves as described.
Invoked in Theorem 11 proof, never constructed.
domain assumption Weighted limits and colimits exist in the dg-categories considered.
Needed for Proposition 4 and Theorem 11.

invented entities (1)

Q, the free closure under perturbations
purpose: To define the dg-functor P and prove functoriality of the perturbation lemma.
Introduced in Theorem 11 without explicit construction or proof of existence.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Enriched categorical aspects of homological perturbation theory." pith.science (2026). https://pith.science/paper/SIOIPBWX

@misc{pith2026241221182,
  author       = {Pith},
  title        = {Pith review of: Enriched categorical aspects of homological perturbation theory},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SIOIPBWX}},
  note         = {Machine review of arXiv:2412.21182}
}

read the original abstract

The paper presents an enriched categorical account of homological perturbation theory, including the formulation, proof and functoriality properties of the homological perturbation lemma.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

7 extracted references · 6 canonical work pages

[1]

Algebraic weak facto risation sys- tems I: Accessible A WFS

Bourke, John, and Richard Garner. “Algebraic weak facto risation sys- tems I: Accessible A WFS.” Journal of Pure and Applied Algebr a 220.1 (2016): 108–147

work page 2016
[2]

Algebraic weak facto risation sys- tems II: Categories of weak maps

Bourke, John, and Richard Garner. “Algebraic weak facto risation sys- tems II: Categories of weak maps.” Journal of Pure and Applie d Algebra 220.1 (2016): 148–174

work page 2016
[3]

Polynomial-time computation of homo topy groups and Postnikov systems in ﬁxed dimension

ˇCadek, Martin, et al. “Polynomial-time computation of homo topy groups and Postnikov systems in ﬁxed dimension.” Siam Journal on Co mputing 43.5 (2014): 1728–1780

work page 2014
[4]

On the perturbation lemma, and deform ations

Crainic, Marius. “On the perturbation lemma, and deform ations.” arXiv preprint math/0403266 (2004). 17

arXiv 2004
[5]

Per- turbation theory in diﬀerential homological algebra II

Gugenheim, Victor KAM, Larry A. Lambe, and James D. Stash eﬀ. “Per- turbation theory in diﬀerential homological algebra II.” Il linois Journal of Mathematics 35.3 (1991): 357–373

work page 1991
[6]

Cau chy complete- ness for DG-categories

Nikoli´ c, Branko, Ross Street, and Giacomo Tendas. “Cau chy complete- ness for DG-categories.” Theory and Applications of Catego ries 37.28 (2021): 940–963

work page 2021
[7]

Homologie des espaces ﬁbr´ es

Shih, Weishu. “Homologie des espaces ﬁbr´ es.” Publicat ions Math´ ematiques de l’IH´ES 13 (1962): 5–87. 18

work page 1962

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [1]

Algebraic weak facto risation sys- tems I: Accessible A WFS

Bourke, John, and Richard Garner. “Algebraic weak facto risation sys- tems I: Accessible A WFS.” Journal of Pure and Applied Algebr a 220.1 (2016): 108–147

work page 2016

[2] [2]

Algebraic weak facto risation sys- tems II: Categories of weak maps

Bourke, John, and Richard Garner. “Algebraic weak facto risation sys- tems II: Categories of weak maps.” Journal of Pure and Applie d Algebra 220.1 (2016): 148–174

work page 2016

[3] [3]

Polynomial-time computation of homo topy groups and Postnikov systems in ﬁxed dimension

ˇCadek, Martin, et al. “Polynomial-time computation of homo topy groups and Postnikov systems in ﬁxed dimension.” Siam Journal on Co mputing 43.5 (2014): 1728–1780

work page 2014

[4] [4]

On the perturbation lemma, and deform ations

Crainic, Marius. “On the perturbation lemma, and deform ations.” arXiv preprint math/0403266 (2004). 17

arXiv 2004

[5] [5]

Per- turbation theory in diﬀerential homological algebra II

Gugenheim, Victor KAM, Larry A. Lambe, and James D. Stash eﬀ. “Per- turbation theory in diﬀerential homological algebra II.” Il linois Journal of Mathematics 35.3 (1991): 357–373

work page 1991

[6] [6]

Cau chy complete- ness for DG-categories

Nikoli´ c, Branko, Ross Street, and Giacomo Tendas. “Cau chy complete- ness for DG-categories.” Theory and Applications of Catego ries 37.28 (2021): 940–963

work page 2021

[7] [7]

Homologie des espaces ﬁbr´ es

Shih, Weishu. “Homologie des espaces ﬁbr´ es.” Publicat ions Math´ ematiques de l’IH´ES 13 (1962): 5–87. 18

work page 1962