Derived mapping spaces of $\infty$-categories

Chris Kapulkin; Daniel Carranza; Kensuke Arakawa

REVIEW 2 major objections 3 minor 10 references

Derived mapping spaces of $\infty$-categories

T0 review · 2 major / 3 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Resolutions compute mapping spaces in localizations.

desk verdict Repackaged lemma, genuinely new cubical applications, with one terse load-bearing appendix step that referees should push on. read the letter →

arxiv 2509.10288 v1 pith:KLMORIXN submitted 2025-09-12 math.AT math.CT

Kensuke Arakawa , Daniel Carranza , Chris Kapulkin This is my paper · ORCID

classification math.ATmath.CT MSC 18N60

keywords derivedmappingspacelemmalocalizationofinfinity-categoriescalculusfractionscubicalcategoriessimplicialhomotopyequivalencesdiscretetheorysets

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 establishes the Derived Mapping Space Lemma: in a relative $\infty$-category $(\mathcal{C},W)$, if an object $Y$ is resolved by a weakly contractible diagram $Y_\bullet$ whose entries are weakly equivalent to $Y$ and whose mapping-space colimit is already local, then the mapping spaces of the localization $\mathcal{C}[W^{-1}]$ are homotopy colimits of the ordinary mapping spaces into the resolution. The lemma is a catch-all statement: it recovers the classical computation of derived mapping spaces by simplicial frames, and it supplies a single sufficient condition under which a cubical or simplicial category is the localization of its underlying category at homotopy equivalences. The condition is checkable directly on the enrichment, and the authors use it to justify that the cubical category of graphs is the localization of the category of graphs at graph homotopy equivalences. A sympathetic reader should care because the lemma turns a difficult question about mapping spaces after localization into a verification performed on a resolution built before localizing.

What carries the argument

The carrying object is the notion of an $I$-resolution (Definition 2.1): a diagram $Y_\bullet:I\to Y\downarrow\mathcal{C}$ such that $I$ is weakly contractible, each structure map $Y\to Y_i$ becomes an equivalence in $\mathcal{C}[W^{-1}]$, and the colimit presheaf $\operatorname{colim}_{i\in I}\mathcal{C}(-,Y_i)$ sends weak equivalences to equivalences of $\infty$-groupoids. The proof mechanism is the adjunction $\gamma_!\dashv\gamma^*$ between presheaf categories: because the colimit presheaf lies in the essential image of $\gamma^*$, the unit at the colimit is an equivalence, and because $I$ is weakly contractible the unit at the constant diagram is an equivalence; comparing the two units in a commutative square forces the map $\Phi_i$ to be an equivalence. Applications lean on weak (co)tensors by the interval cube $\square^1$, representatives of the tensor/cotensor adjunction with only unenriched naturality, a flexibility needed because the cubical geometric product is not symmetric; these build resolutions whose mapping-space colimit is the original mapping space.

What would settle it

Concretely, the theorem would be falsified by any relative $\infty$-category and resolution satisfying (R1)-(R3) for which $\gamma^*\mathcal{C}[W^{-1}](-,Y)$ and $\operatorname{colim}_{i\in I}\mathcal{C}(-,Y_i)$ are not equivalent as presheaves. A reader can search for such an example by taking the paper's own counterexample of a one-object cubical category built from a cubical group with nontrivial higher homotopy and trying to equip it with a resolution that satisfies (R3); the lemma says no such resolution exists, so exhibiting one would be a direct counterexample, while proving none exists would confirm the sharp role of (R3).

Watch

Extended reading notes

Core claim

The paper's central claim is that derived mapping spaces are computed by resolutions before localization. Concretely, for a relative $\infty$-category $(\mathcal{C},W)$ and an $I$-resolution $Y_\bullet:I\to Y\downarrow\mathcal{C}$ of $Y$, Theorem 2.2 gives an equivalence of presheaves $\gamma^*\mathcal{C}[W^{-1}](-,Y)\simeq\operatorname{colim}_{i\in I}\mathcal{C}(-,Y_i)$, hence equivalences $\mathcal{C}[W^{-1}](X,Y)\simeq\operatorname{hocolim}_{i\in I}\mathcal{C}(X,Y_i)$ natural in $X$. The resolution is not required to be a fibrant replacement in a model category: it only needs a weakly contractible indexing category, weak equivalences from $Y$ to every $Y_i$, and locality of the colimit presheaf. The paper then proves that this single lemma instantiates to simplicial frames in model categories, to a localization criterion for cubical, simplicial, and 2-categories, and to the cubical category of graphs.

Load-bearing premise

The proof hangs on condition (R3): the colimit of mapping spaces into the resolution must already send weak equivalences of $\mathcal{C}$ to equivalences of $\infty$-groupoids; if only the other two conditions hold, the key comparison map is not known to be an equivalence and the theorem does not apply.

Editorial extensions

If this is right

For every model category $\mathcal{M}$, each simplicial resolution $Y_\bullet$ of $Y$ gives $\mathcal{M}[W^{-1}](X,Y)\simeq\mathcal{M}(X,Y_\bullet)$ for cofibrant $X$, recovering frame computations of derived mapping spaces from a single lemma.
Any cubical category admitting weak tensors or weak cotensors by $\square^1$ becomes, after fibrant replacement in the model structure on cubical categories, the localization of its underlying category at homotopy equivalences; the criterion transfers to simplicial categories and to 2-categories with groupoid mapping categories.
The cubical category of graphs is the localization of the category of graphs at graph homotopy equivalences, closing the gap in the previously announced resolution of an open problem in discrete homotopy theory.
Because the associated quasicategory of the cubical category of graphs lacks pushouts, graph homotopy equivalences cannot be the weak equivalences of any model structure on graphs.
Any colimit of representables along a resolution that is already local must agree with the derived representable, so local presheaves can be recognized resolution-by-resolution rather than only by fibrant replacement.

Reading between the lines

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

Beyond the paper: the lemma is really a criterion for when the homotopy colimit of ordinary representables over a resolution is the derived representable, which suggests a descent-style characterization of local presheaves on a relative $\infty$-category.
Beyond the paper: since only weak tensors or cotensors by an interval object are needed, the localization criterion should extend to enrichments over any monoidal category with a chosen interval object, even when the monoidal product is not symmetric; testing this on other categories with box products is a natural next step.
Beyond the paper: the graph result indicates a broader phenomenon in discrete homotopy theory, where concrete categories whose homotopy relation comes from an interval graph may all admit cubical enrichments of this type, and the same theorem would identify their localizations without constructing a model structure.

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

Punchline: the derived mapping space lemma is a repackaging of Cisinski's calculus-of-fractions theorem — the authors admit as much in Remark 2.4 — but the paper earns its keep on the application side. The cubical localization criterion and the graph corollary are genuinely new and close a real gap in discrete homotopy theory.

The proof of Theorem 2.2 is short and correct. The crucial hypothesis is (R3), and the authors do not paper over it: it is exactly the condition that the colimit presheaf is already local, and in each application they verify it by comparing the colimit to an ordinary mapping space. I checked the tensor half of Theorem 3.8 and the stress-tester is right that the terminal object in □ makes it collapse to C0(-,Y), so no hidden asymmetry. The reductions for the simplicial and 2-categorical corollaries are believable, and the graph application is a nice payoff.

The soft spot is Appendix B, specifically Corollary B.21. The proof of item (2) replaces a locally Kan cubical category C by U•Ex∞•T•C in one sentence, and that replacement is load-bearing for identifying the mapping spaces of the homotopy-coherent nerve. It may be true — the surrounding statements are cited correctly — but as written it reads like a handwave. A referee should push for a more detailed argument or a precise reference. The cubical Bergner model structure also appears with a proof assembled from Lurie and Stanculescu; that is fine, but the paper is not self-contained in a way that makes the appendix easy to check.

No circularity. The disclosure that the main lemma follows from Cisinski's theorem is a point in the authors' favor, not against them. The citation pattern is appropriate — they cite their own prior work where it is actually prior work.

Who should read this: anyone working on mapping spaces in localizations, cubical models of higher categories, or discrete homotopy theory. It deserves a serious referee. I would send it to review and ask for the appendix to be expanded before publication.

Referee Report

2 major / 3 minor

Summary. The paper proves a Derived Mapping Space Lemma (Theorem 2.2) which gives conditions under which the mapping space in a localization C[W^{-1}] can be computed as a colimit of mapping spaces into a resolution Y• of Y before localizing. The lemma is applied in three directions: it recovers the model-categorical frame computation of mapping spaces (Corollary 3.1); it gives a sufficient condition for a cubical category to be the localization of its underlying 1-category at homotopy equivalences (Theorem 3.8), with a simplicial analogue and a 2-categorical specialization (Corollaries 3.15 and 3.16); and it proves that the cubical category of graphs is the localization of the category of graphs at graph homotopy equivalences (Corollary 3.18). The paper also develops, in appendices, background on cubical sets and cubical categories, including a cubical Bergner model structure, a Quillen equivalence with the Joyal model structure, and a comparison of mapping spaces under the cubical homotopy-coherent nerve.

Significance. The Derived Mapping Space Lemma is a concise and general statement that unifies several known localization techniques and supplies a new proof distinct from Cisinski's calculus of fractions. If the applications are fully established, the paper provides a valuable framework: the cubical graph application (Corollary 3.18) closes a known gap in discrete homotopy theory, and the cubical results are genuinely not obtainable by a purely simplicial argument. The proof of Theorem 2.2 itself is clean and the hypotheses (R1)-(R3) are checkable in the examples. However, the paper's utility is contingent on the correctness of Theorem 3.8 for both weak tensors and weak cotensors, and on the completeness of the Appendix B comparison result; both currently need further work.

major comments (2)

[Theorem 3.8, proof] The sentence 'We prove the result for weak cotensors; the case of weak tensors proceeds analogously' is not justified and appears to be incorrect as written. For weak cotensors, the resolution is the functor -⋔Y : □^op → Y↓C0; since □^op has an initial object, the homotopy colimit over □^op does not collapse, and Corollary 3.14 identifies it with the full cubical mapping space C(X,Y), giving condition (R3). For weak tensors, the natural functor is Y⊗- : □ → Y↓C0, and □ has a terminal object [1]^0. The homotopy colimit over □ therefore collapses to the value at the terminal object, so the colimit presheaf is C0(-,Y), regarded as a discrete ∞-groupoid-valued presheaf. Condition (R3) for this discrete presheaf would require that precomposition with every homotopy equivalence induces a bijection on the underlying sets of morphisms, which is not implied by the homotopy-invariance of the cubical mapping spaces and fails in general. Thus the weak-tensor half of Theorem 3.8 is not established by the given argument, and the theorem as stated may overreach. Please either supply a correct proof for the weak-tensor case or restrict the statement (and the dependent corollaries) to weak cotensors.
[Appendix B, Corollary B.21] The proof of Corollary B.21 is only a sketch. The central reduction, 'Using Proposition B.20, we can replace C appropriately (e.g. by U•Ex∞•T•C) and assume that C = U•D', is asserted without justification. It must be shown that this replacement is a DK-equivalence, that applying Ex∞ levelwise to the hom-objects of T•C actually produces a simplicial category (i.e., that the enrichment structure is preserved, which is not automatic because Ex∞ is not visibly lax monoidal), and that the resulting equivalences of mapping spaces are compatible with the representables of N□C. Corollary B.21 is load-bearing: it is used in the proof of Theorem 3.8 to identify N□Cf(-,Y) with Cf(-,Y). Without a complete and detailed proof of this corollary, the localization theorem for cubical categories is not fully supported.

minor comments (3)

[Theorem 3.8, proof] In the final paragraph of the proof, the names β, φ, and Φ appear to be interchanged: the map β, whose domain is C0[W^{-1}](-,Y), should be identified with Φ_[1]^0 from diagram (∗), while the dotted arrow from C0(-,Y) should be identified with φ_[1]^0. As written, the sentence 'we can identify the map β as the map φ_[1]^0' is type-incorrect and makes the argument difficult to follow.
[Corollary 3.1 and Corollary 3.14] The statements 'every simplicial set K is its own geometric realization, i.e. colim_{Δ^op} K → Set → S ≃ K' and 'X is the (homotopy) colimit of the diagram □^op → Set → S' are imprecise: the colimit must be understood as a homotopy colimit of the diagram of discrete spaces, not as a strict 1-categorical colimit, since the latter would give the set of 0-cubes (because Δ^op and □^op have initial objects). A short clarifying remark would prevent confusion.
[Acknowledgments] There is a typo in 'D.C. acknolwedges' (should be 'acknowledges').

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the Derived Mapping Space Lemma is a genuine theorem with a verified hypothesis, and applications check the hypothesis rather than assuming the conclusion.

full rationale

The paper's main theorem (Theorem 2.2) is not circular: its proof uses conditions (R1)-(R3) exactly as stated. (R1) makes the indexing category weakly contractible, (R2) ensures the localized structure maps are equivalences, and (R3) is used only to ensure the colimit presheaf lies in the essential image of gamma^*, so the right-hand unit in diagram (∗) is an equivalence. The conclusion is not assumed in (R3); rather, (R3) is a genuine hypothesis that would indeed be a consequence of the conclusion, but the proof never invokes the conclusion to establish it. In applications, (R3) is verified independently: in Theorem 3.8 it is checked via the equivalence colim C0(X,□^n ⋔ Y) ≃ C(X,Y) (Corollary 3.14) together with the fact that precomposition with a homotopy equivalence is itself a homotopy equivalence, which follows from the consequences of Definition 3.2. The result for graphs (Corollary 3.18) uses [CK24, Thm. 4.1] for Kan-ness of N^G and levelwise weak equivalences; that is prior work establishing cubical nerve properties, not the localization statement itself, so it is external support rather than a self-referential reduction. Appendix results are either proved in the paper or cited to Lurie, Cisinski, KV20, and Stănculescu; none of these import the main theorem. Remark 2.4's admission that Theorem 2.2 can be reduced to Cisinski's Theorem 7.2.8 reduces novelty but is an external benchmark, not circularity. No fitted parameter is renamed as a prediction and no equation reduces to its inputs by construction.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The central lemma rests on standard infinity-category and model-category machinery. The most paper-specific inputs are the cubical Bergner model structure and the mapping-space comparison in Appendix B, both proven via cited results, and the graph-nerve theorem of [CK24]. No free parameters are fitted.

assumptions (6)

standard math Localizations of relative infinity-categories exist and are unique up to contractible choice [Lur25, tags 01N0, 01N1].
Invoked in Definition 1.1 to define C[W^{-1}] and used throughout the paper.
standard math For a weakly contractible infinity-category I, the constant diagram at X is a homotopy colimit cone [Lur09, Corollary 4.4.4.10].
Used in the proof of Theorem 2.2 and to verify condition R1 in applications.
standard math Fully faithful functors reflect equivalences.
Used implicitly when deducing Φ_i is an equivalence from γ_!φ_i being an equivalence.
standard math The cubical Bergner model structure on cubical categories exists and is Quillen equivalent to the Bergner model structure on simplicial categories.
Proven in Appendix B using [Lur09, Prop. A.3.2.4] and [Sta14, Prop. 2.3]; load-bearing for the cubical localization theorem.
domain assumption Mapping spaces of a locally Kan cubical category agree with mapping spaces of its cubical homotopy coherent nerve.
Corollary B.21 is used in the proof of Theorem 3.8 to identify mapping spaces of N□Cf with colimits of cubical mapping spaces.
domain assumption The m-nerves N^G_m: Graph → cSet are lax monoidal, take values in cubical Kan complexes, and the inclusions N^G_m → N^G_{m+1} are weak equivalences [CK24, Thm. 4.1].
Used in Corollary 3.18 to promote the category of graphs to a cubical category and to identify Graph1 → Graph∞ as a weak equivalence.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Derived mapping spaces of $\infty$-categories." pith.science (2026). https://pith.science/paper/KLMORIXN

@misc{pith2026250910288,
  author       = {Pith},
  title        = {Pith review of: Derived mapping spaces of $\infty$-categories},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KLMORIXN}},
  note         = {Machine review of arXiv:2509.10288}
}

abstract

We prove the Derived Mapping Space Lemma, which generalizes the central theorem of Cisinski's work on calculus of fractions for $\infty$-categories, and allows us to provide a unified framework for analyzing mapping spaces in localizations of ($\infty$-)categories. As an application, we give a sufficient condition for when a cubical or simplicial category is the localization of its underlying category at homotopy equivalences.

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

10 extracted references · 10 canonical work pages

[1]

[Ber07] J. E. Bergner,A model category structure on the category of simplicial categories, Trans. Amer. Math. Soc.359(2007), no. 5, 2043–2058. [BL05] H. Barcelo and R. Laubenbacher,Perspectives on𝐴-homotopy theory and its applications, Discrete Math.298(2005), no. 1-3, 39–61. [Cis06] D.-C. Cisinski,Les pr ´efaisceaux comme mod`eles des types d’homotopie, ...

work page 2007
[326]

Kapulkin and N

[KK24] K. Kapulkin and N. Kershaw,Closed symmetric monoidal structures on the category of graphs, Theory Appl. Categ.41(2024), Paper No. 23, 760–784. [KK25] K. Kapulkin and N. Kershaw,Data analysis using discrete cubical homology,

work page 2024
[1967]

Hinich,Dwyer-Kan localization revisited, Homology Homotopy Appl.18(2016), no

[Hin16] V. Hinich,Dwyer-Kan localization revisited, Homology Homotopy Appl.18(2016), no. 1, 27–48. [Hir03] P. S. Hirschhorn,Model categories and their localizations, Mathematical Surveys and Monographs, vol. 99, American Mathematical Society, Providence, RI,

work page 2016
[1999]

[Jar06] J. F. Jardine,Categorical homotopy theory, Homology Homotopy Appl.8(2006), no. 1, 71–144. [Joy09] A. Joyal,The theory of quasi-categories and its applications,

work page 2006
[2003]

[Hov99] M

errata available athttps://math.mit.edu/psh/MCATL-errata-2018-08-01.pdf. [Hov99] M. Hovey,Model categories, Mathematical Surveys and Monographs, vol. 63, American Mathematical Society, Providence, RI,

work page 2018
[2008]

Joyal and M

[JT07] A. Joyal and M. Tierney,Quasi-categories vs Segal spaces, Categories in algebra, geometry and mathematical physics, 2007, pp. 277–

work page 2007
[2014]

Rivera and M

[RZ18] M. Rivera and M. Zeinalian,Cubical rigidification, the cobar construction and the based loop space, Algebr. Geom. Topol.18(2018), no. 7, 3789–3820. [Sta14] A. E. Stanculescu,Constructing model categories with prescribed fibrant objects, Theory Appl. Categ.29(2014), No. 23, 635–653. 26

work page 2018
[2019]

Carranza and K

[CK24] D. Carranza and K. Kapulkin,Cubical setting for discrete homotopy theory, revisited, Compos. Math.160(2024), no. 12, 2856–2903. [CKK23] D. Carranza, K. Kapulkin, and J. Kim,Nonexistence of colimits in naive discrete homotopy theory, Appl. Categ. Structures31(2023), no. 5, Paper No. 41,

work page 2024

Show all 10 references

[2023]

[CKW25] D

preprint. [CKW25] D. Carranza, K. Kapulkin, and L. Z. Wong,Diagonal lemma for presheaves on elegant Reedy categories, Theory Appl. Categ.44(2025), Paper No. 11, 326–343. MR4901943 [DKLS24] B. Doherty, K. Kapulkin, Z. Lindsey, and C. Sattler,Cubical models of(∞,1)-categories, M...

2025
[2025]

[KV20] K

preprint. [KV20] K. Kapulkin and V. Voevodsky,A cubical approach to straightening, J. Topol.13(2020), no. 4, 1682–1700. [Lur09] J. Lurie,Higher topos theory, Annals of Mathematics Studies, vol. 170, Princeton University Press, Princeton, NJ,

2020

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.

[1] [1]

[Ber07] J. E. Bergner,A model category structure on the category of simplicial categories, Trans. Amer. Math. Soc.359(2007), no. 5, 2043–2058. [BL05] H. Barcelo and R. Laubenbacher,Perspectives on𝐴-homotopy theory and its applications, Discrete Math.298(2005), no. 1-3, 39–61. [Cis06] D.-C. Cisinski,Les pr ´efaisceaux comme mod`eles des types d’homotopie, ...

work page 2007

[2] [326]

Kapulkin and N

[KK24] K. Kapulkin and N. Kershaw,Closed symmetric monoidal structures on the category of graphs, Theory Appl. Categ.41(2024), Paper No. 23, 760–784. [KK25] K. Kapulkin and N. Kershaw,Data analysis using discrete cubical homology,

work page 2024

[3] [1967]

Hinich,Dwyer-Kan localization revisited, Homology Homotopy Appl.18(2016), no

[Hin16] V. Hinich,Dwyer-Kan localization revisited, Homology Homotopy Appl.18(2016), no. 1, 27–48. [Hir03] P. S. Hirschhorn,Model categories and their localizations, Mathematical Surveys and Monographs, vol. 99, American Mathematical Society, Providence, RI,

work page 2016

[4] [1999]

[Jar06] J. F. Jardine,Categorical homotopy theory, Homology Homotopy Appl.8(2006), no. 1, 71–144. [Joy09] A. Joyal,The theory of quasi-categories and its applications,

work page 2006

[5] [2003]

[Hov99] M

errata available athttps://math.mit.edu/psh/MCATL-errata-2018-08-01.pdf. [Hov99] M. Hovey,Model categories, Mathematical Surveys and Monographs, vol. 63, American Mathematical Society, Providence, RI,

work page 2018

[6] [2008]

Joyal and M

[JT07] A. Joyal and M. Tierney,Quasi-categories vs Segal spaces, Categories in algebra, geometry and mathematical physics, 2007, pp. 277–

work page 2007

[7] [2014]

Rivera and M

[RZ18] M. Rivera and M. Zeinalian,Cubical rigidification, the cobar construction and the based loop space, Algebr. Geom. Topol.18(2018), no. 7, 3789–3820. [Sta14] A. E. Stanculescu,Constructing model categories with prescribed fibrant objects, Theory Appl. Categ.29(2014), No. 23, 635–653. 26

work page 2018

[8] [2019]

Carranza and K

[CK24] D. Carranza and K. Kapulkin,Cubical setting for discrete homotopy theory, revisited, Compos. Math.160(2024), no. 12, 2856–2903. [CKK23] D. Carranza, K. Kapulkin, and J. Kim,Nonexistence of colimits in naive discrete homotopy theory, Appl. Categ. Structures31(2023), no. 5, Paper No. 41,

work page 2024