2-dimensional Lawvere theories, commutativity, and higher Day convolution

Tom\'a\v{s} Perutka

REVIEW 2 major objections 3 minor 1 cited by

A lax commutativity structure on a Lawvere 2-theory turns the 2-category of models and lax maps into a closed 2-multicategory, yielding a comonad on models that generalizes Fox's theorem.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-02 23:14 UTC pith:TCJKOSJJ

load-bearing objection A genuinely new 2-dimensional framework, with the main closed-structure theorem resting on a self-declared Fubini gap that is plausibly patched but not yet independently checked. the 2 major comments →

arxiv 2602.14332 v2 pith:TCJKOSJJ submitted 2026-02-15 math.CT math.ATmath.LO

2-dimensional Lawvere theories, commutativity, and higher Day convolution

Tom\'a\v{s} Perutka This is my paper

classification math.CT math.ATmath.LO MSC 18D0518N1018D20

keywords Lawvere 2-theories2-dimensional commutativitylax endsclosed 2-multicategoriesinternal homDay convolutionFox theoremEckmann-Hilton argument

verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper sets up a 2-dimensional analogue of commutative algebraic theories: a notion of w-commutativity for Lawvere 2-theories, for w ∈ {strict, pseudo, lax, colax}. Its central claim is that whenever a Lawvere 2-theory T carries such a commutativity structure, the 2-category of T-models and lax homomorphisms naturally forms a closed 2-multicategory: for any two models X,Y there is an internal-hom model Hom_l(X,Y), whose underlying category is the category of lax homomorphisms from X to Y. This yields a general mechanism for internalizing operations: the category of internal algebras IntAlg(X) of any model X again carries a T-model structure, and the assignment X ↦ IntAlg(X) is a comonad, idempotent under 2-dimensional Eckmann-Hilton conditions. The paper also obtains a generalized Day convolution: for any monoidal 2-category V and monoids M,N in it, the hom-category V(M,N) carries a multicategory structure. Together these results generalize Fox's theorem about commutative monoids to arbitrary lax commutative theories, and the formalism is designed to extend to (∞,2)-categories.

Core claim

Let T be a Lawvere 2-theory equipped with a lax commutativity, i.e. a pseudomonoid structure µ:T⊗_{s,l}T→T in the monoidal 2-category of F-sketches. The main theorem (Theorem 2) states that the 2-multicategory Mod_l(T,Cat) is closed: for models X,Y,Z there is a natural isomorphism Mul_2(X,Y;Z) ≅ Hom_l(X, Hom_l(Y,Z)), where Hom_l(Y,Z) is the T-model defined by Hom_l(Y,Z)(n)=Hom_l(Y,Z)^n with T-action induced by the lifted model eY. In particular the 2-functor IntAlg := Hom_l(∗,−) promotes to an endo-2-functor on Mod_l(T,Cat). As a formal consequence (Theorem 3), IntAlg is a comonad, and it is idempotent whenever T is Eckmann-Hilton; this recovers and extends Fox's idempotent comonad theorem.

What carries the argument

The central object is the w-commutativity structure: a pseudomonoid structure µ:T⊗_{s,w}T→T on the Lawvere 2-theory T, taken in the monoidal 2-category of F-sketches with the semi-strict Gray tensor product ⊗_{s,w}. This structure produces, for each model X, a lifted model eX:T→Mod_w(T,C), whose components are the w-homomorphisms X(β):X^k→X^l; the internal hom Hom_w(X,Y) is then built from these lifted maps by postcomposition. The load-bearing technical device is the generalized Fubini rule for lax ends (Prop. A.7), which allows an iterated lax end over T⊗_{s,l}T to be reindexed as a double lax end.

Load-bearing premise

The entire proof of the closed structure rests on a new Fubini rule for lax ends, Proposition A.7: that ∫^lax_{a,b} T(a,b,a,b) ≅ ∫^lax_a ∫^lax_b T(a,b,a,b) for T:A^op×B^op×(A⊗_lB)→Cat; if this reindexing fails for some T, the isomorphism Mul_2 ≅ Hom_l(−,Hom_l(−,−)) is not established.

What would settle it

Check the comparison map in Proposition A.7 for A = the walking 2-cell (two objects, one non-invertible 2-cell) and a 2-functor T:A^op×A^op×(A⊗_l A)→Cat that records the lax Gray cell Σ_{f,g}. If the iterated lax end and the simultaneous lax end are not isomorphic, the proof of Theorem 2 collapses at line (11); agreement in such small test cases would support the closed-structure claim.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

For every lax commutative theory T, Hom_l(X,Y) is itself a T-model, so the hom-category of lax maps between any two models carries the same algebraic structure as the models themselves.
IntAlg becomes an endo-2-functor and a comonad on Mod_l(T,Cat); when T is Eckmann-Hilton the comonad is idempotent, so internal algebras form a property-like structure on models.
The generalized Day convolution (Lemma 4.5) puts a multicategory structure on V(M,N) for monoids M,N in any monoidal 2-category, without requiring left Kan extensions; when V has them, the structure is representable, recovering ordinary Day convolution as the case V=Cat, M=1, N=presheaves.
The Eckmann-Hilton conditions (no active unary maps, unital active maps, invertible diagonal cells) make the lifting functor Mod_l(T,C)→Mod_l(T,Mod_l(T,C)) an equivalence, extending the classical Eckmann-Hilton argument to theories.
The author argues the results are formal enough that the definitions and Theorem 1 carry to Lawvere (∞,2)-theories; Theorem 2 awaits an ∞-categorical version of the lax-end Fubini rule.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The closed 2-multicategory structure suggests a semantic characterization of higher commutativity: a theory is w-commutative exactly when its model 2-category is closed in this sense — the paper proves the forward direction explicitly, and Theorem 1's syntax-semantics bijection points toward the converse.
The idempotent comonad IntAlg may define a general notion of 'cartesian-like' models for any Eckmann-Hilton theory: by analogy with Fox's theorem, its coalgebras should be precisely the T-models with finite lax/oplax products — a characterization the paper leaves open.
The generalized Fubini lemma, if reformulated via F-category structures as 'partially lax limits' over twisted arrows (the direction hinted at in Remark A.1), would likely make the entire proof formal enough to transfer verbatim to (∞,2)-categories, where the explicit diagram chase does not apply.
The multicategory on V(M,N) exists without Kan extensions, hinting that Day convolution is not a property of the ambient category but a structure on hom-categories of monoids; this may offer a route to convolution-like tensor products in monoidal 2-categories lacking left Kan extensions.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Desk Editor's Note

A genuinely new 2-dimensional framework, with the main closed-structure theorem resting on a self-declared Fubini gap that is plausibly patched but not yet independently checked.

read the letter

Dear colleague,

This is a real paper with real new results. The lax-case closed 2-multicategory structure (Thm 2), the partial Fox comonad (Thm 3), miracle associativity (Prop 3.7), and Day convolution without Kan extensions (Prop C/Lemma 4.5) are all new as far as I can tell. The paper is also unusually honest: Remark 4.9 flags the non-standard Fubini step, Section 5.3 says the ∞-analogue of Thm 2 is unproved, and Section 6 admits the coalgebra half of Fox is open. That transparency earns credit.

The soft spot is exactly where the paper points. Theorem 2 is proved by one chain of lax-end isomorphisms, and the only non-routine step is the iterated lax end. Appendix A patches it with Prop A.7, but that lemma is new, diagram-heavy, and leaves some steps as “clearly a bijection.” I checked the stress-test worry about Lemma A.6: on my reading it does not land. In the semi-strict tensor, a Gray cell with a tight coordinate is forced to be an identity; on the other side of the claimed equivalence, the corresponding Gray cell in A⊗_l B is also an identity when one coordinate is the identity. So the two sides do match. But that does not make the appendix easy: Prop A.7 is still a bespoke Fubini rule for lax ends, not a variant of a known result, and it depends on a custom F-category structure plus [Hir22].

I would send this to a serious referee. The framework is coherent, the theorems are derived rather than fitted, and the examples (Schmitt, Day convolution, the T_braid failure) give real external checks. The right referee is someone who can verify Appendix A by hand. If Prop A.7 holds, the paper is a solid contribution; if it fails, Thm 2 collapses, though the remainder still has value.

Who is this for? 2-category theorists working on multicategories, monads, and higher commutativity. It deserves a serious referee, and the request should be to expand the compressed steps in the appendix and ideally find a more conceptual proof of the Fubini rule.

Best, [Your name]

Referee Report

2 major / 3 minor

Summary. The paper develops a 2-dimensional theory of commutativity for Lawvere 2-theories, formalized via pseudomonoid structures in the monoidal 2-category of F-sketches with semi-strict Gray tensor products. It states three main results: a syntax-semantics comparison for w-commutativity (Theorem 1), a closed 2-multicategory structure on Mod_w(T,Cat) for a lax-commutative Lawvere 2-theory (Theorem 2), and a partial Fox theorem giving an idempotent comonad IntAlg under Eckmann–Hilton conditions (Theorem 3). The proof of Theorem 2 is by lax ends; the critical Fubini step is explicitly flagged as nonstandard in Remark 4.9 and deferred to Appendix A. There is also a generalized Day convolution multicategory for hom-categories between monoids in a monoidal 2-category, and an outline of an (∞,2)-categorical analogue.

Significance. If Theorem 2 is correct, it is a substantial result: it would provide internal homs for lax model categories of any lax-commutative Lawvere 2-theory, and via Theorem 3 an idempotent comonad generalizing Fox's theorem. The Day-convolution construction (Lemma 4.5 / §5.2) is of independent interest, and the paper is commendably explicit about where its proof is incomplete. However, the central closed-structure claim is currently not established: the appendix's Fubini rule is stated for a different tensor product than the one used in Theorem 2, and the key verification contains asserted bijections. The paper's other contributions do not by themselves discharge the main theorem.

major comments (2)

[§4.3, Theorem 2, Remark 4.9, and Appendix A] The Fubini step in Eq. (11) is the load-bearing step of Theorem 2. Remark 4.9 explicitly states that the usual Fubini rule does not apply because the wedge is indexed by T^op×T^op×(T⊗_{s,l}T), not by T^op×T^op×T×T. The promised replacement, Prop. A.7, is stated for a functor A^op×B^op×(A⊗_lB)→Cat, i.e. with the fully lax Gray tensor product A⊗_lB. These are not the same tensor product when T is a Lawvere 2-theory: in T⊗_{s,l}T the Gray cells Σ_{αβ} are identities when α or β is tight (inert), whereas in T⊗_lT they are generally non-invertible lax cells. Thus Prop. A.7 does not, on its face, apply to the wedge appearing in line (11). The author must either prove the semi-strict version of Prop. A.7 or give a precise reduction from the semi-strict case to the fully lax case.
[Appendix A, Lemma A.6 and Prop. A.7] Even granting the fully lax version, the proof of Prop. A.7 is not complete. The claimed equivalence j in Lemma A.6 is verified only by saying that nontrivial Gray cells match; the cases where one of the two 1-cells is tight are not checked, and those are exactly the cases where the left-hand semi-strict tensor product imposes identities. The final step of Prop. A.7 asserts that the correspondence is 'clearly a bijection', but that correspondence is precisely where the Gray cells Σ_{f,g} must be shown to satisfy the lax-wedge equations (16)–(17). The proof also says most verifications follow as in [Hir22], but the whole point of the appendix is that the ordinary Fubini rule of Lemma A.4 does not cover the present wedge. As written, this leaves the central isomorphism (11) without a certified proof.

minor comments (3)

[§4.4, Theorem 3] The proof says 'From that, it formally follows' that Hom_w(∗,−) is a comonad, with a footnote crediting J. Bourke. Since the rest of the paper is very detailed, this formal fact should be stated as a lemma with at least a sketch, especially because Theorem 3 is one of the main advertised results.
[§3.2, Example 3.12] The claim that Proposition 3.7 yields exactly the Yang–Baxter equation is left as an exercise. A short explicit display would help the reader verify this important motivating example.
[§5.3] The discussion of the (∞,2)-categorical analogue says the final question 'can we adapt the proof of Theorem 2' would require ∞-categorical lax ends that are not developed. In view of the gap in the ordinary proof, this should be phrased as a conjecture/open problem rather than as a routine extension.

Circularity Check

0 steps flagged

No circularity: Theorem 2 is a genuine derivation, and the flagged Fubini step is a technical verification gap, not an input-output identity.

full rationale

Walking the derivation chain, I find no circular step. Theorem 2's closed structure is proved by a chain of lax-end isomorphisms starting from the definitions of models and the lax Gray tensor product; the internal hom Hom_l(X,Y) is independently constructed in Definition 3.22 from the lift eY, and the isomorphism Mul_2(X,Y;Z) ≅ Hom_l(X, Hom_l(Y,Z)) is then derived rather than assumed. The single delicate step, the Fubini isomorphism (line (11)), is explicitly flagged in Remark 4.9: "The gap in the proof above is that the usual Fubini rule cannot be applied." It is addressed by a new technical lemma, Proposition A.7, in Appendix A. This lemma is not an input of Theorem 2; it is a separate result about lax ends. Its proof is compressed and relies on diagram chases and some claims such as "this correspondence is clearly a bijection," which the paper itself acknowledges in Appendix A is delicate and not suitable for naive ∞-categorical generalization. That is a verification/correctness risk, not circularity. There are no fitted parameters or data being renamed as predictions. The external citations ([ABK24], [Hir22], [HP02], [Ber20], [Fox76]) are to work by other authors, and there are no self-citations to the present author's prior results forming a load-bearing chain. The footnote about John Bourke's explanation of the formal comonad step is an omitted proof, not an appeal to the paper's own authority. Theorem 1 is a Yoneda-style bijection and is close to definitional in character, but it does not assume Theorem 2 and does not force the closed-structure conclusion. The central claims retain independent mathematical content, so the appropriate circularity score is 0.

Axiom & Free-Parameter Ledger

0 free parameters · 5 axioms · 0 invented entities

The central claims rest on two imported frameworks taken as black boxes: the F-sketches/enhanced 2-category machinery of [ABK24] (§2.4–2.5) and the lax-end calculus of [Hir22] (Theorem 2; patched by Appendix A). No numbers are fitted anywhere; the Eckmann–Hilton conditions (Def 3.28) are explicit hypotheses, not hidden parameters. The (∞,2)-section additionally depends on a same-month preprint [AHM26]. New structures (w-commutativity, the Hom^w_V multicategory) are defined in the text and their properties proven, so no separately postulated entity lacks support.

axioms (5)

domain assumption Semi-strict Gray tensor products ⊗_{w',w} of F-categories/F-sketches exist with the universal properties of [ABK24, Thm 6.12, Prop 7.1, §6.15].
Imported without re-proof (§2.4–2.5). The definition of w-commutativity (Def 3.1) is a pseudomonoid in (F-Sk_{s,w}, ⊗_{s,w}, 1), so the whole paper rests on this framework.
domain assumption Lax ends compute lax natural transformations, with the Fubini-style results of [Hir22, Prop 3.2, Prop 3.3].
Used in line (12) of Theorem 2's proof; Appendix A extends [Hir22] and admits (Remark A.1) that lax ends lack the ∞-categorical analogue needed for Section 5.
standard math Uniqueness of the magmal structure on F^op (Lemma 1.13; one step dualized from [Ark25, Thm 1.1]).
Fixes µ(m,n) = m·n and the unit u(∗)=1 used to unwrap Def 3.1.
domain assumption Marked (∞,2)-category results of [AGH24, 2.8.1] and the contemporaneous [AHM26, 2.3].
Load-bearing for Section 5 only; the author flags [AHM26] as established "very recently" (§5.1).
standard math Unbiased pseudomonoid definitions and coherences [Lei04, §3.1].
Used to define the multicategory on V(M1,M2) in Def 4.1 and Lemma 4.5.

pith-pipeline@v1.3.0-alltime-deepseek · 43737 in / 20924 out tokens · 190072 ms · 2026-08-02T23:14:30.245309+00:00 · methodology

0 comments

read the original abstract

The aim of this paper is to study categorified algebraic structures and their pseudo- and lax homomorphisms using the framework of Lawvere $2$-theories, and more generally, (enhanced) $2$-dimensional sketches. The key notion we focus on is that of $2$-dimensional commutativity. As one of the main results, we prove that if a Lawvere $2$-theory $\mathbb{T}$ is equipped with such a structure, then the $2$-category $\mathsf{Mod}_l(\mathbb{T},\mathbf{Cat})$ of $\mathbb{T}$-models, lax homomorphisms, and modifications admits a natural structure of a closed $2$-multicategory. From this, we deduce a generalization of Fox's theorem. We also discuss the analogue in the higher setting for Lawvere $(\infty,2)$-theories. As a result of independent interest, we construct a multicategory (or $\infty$-operad) structure on the hom-category $\mathsf{Hom}_{\mathbb{V}}(\mathcal{M},\mathcal{N})$, where $\mathbb{V}$ is a monoidal $(\infty,2)$-category and $\mathcal{M},\mathcal{N}$ are monoids therein.

discussion (0)

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

Controlled theories, categorification, and homotopification
math.CT 2026-07 conditional novelty 7.0

Controlled theories yield functorial Lawvere 2-theories and simplicial Lawvere theories, producing a new model of ∞-groups and a candidate for infinite loop spaces.

Reference graph

Works this paper leans on

10 extracted references · 7 linked inside Pith · cited by 1 Pith paper

[1]

Enhanced 2-categorical structures, two- dimensional limit sketches and the symmetry of internalisation.arXiv:2412.07475,

[ABK24] Nathanael Arkor, John Bourke, and Joanna Ko. Enhanced 2-categorical structures, two- dimensional limit sketches and the symmetry of internalisation.arXiv:2412.07475,

Pith/arXiv arXiv
[3]

Free fibrations, lax colimits and kan extensions (∞,2)-categories.arXiv:2602.07604,

[AHM26] Fernando Abell´ an, Rune Haugseng, and Louis Martini. Free fibrations, lax colimits and kan extensions (∞,2)-categories.arXiv:2602.07604,

arXiv
[1999]

Tensor product for symmetric monoidal categories.arXiv:0711.0324,

[Sch07] Vincent Schmitt. Tensor product for symmetric monoidal categories.arXiv:0711.0324,

Pith/arXiv arXiv
[2010]

Nuclei of categories with tensor products.arXiv:0708.2761,

[Dav07] Alexei Davydov. Nuclei of categories with tensor products.arXiv:0708.2761,

Pith/arXiv arXiv
[2012]

Linearly distributive fox theorem.arXiv:2506.02180,

64 [KB25] Rose Kudzman-Blais. Linearly distributive fox theorem.arXiv:2506.02180,

arXiv
[2013]

Notes on lax ends.arXiv:2210.01522,

[Hir22] Kengo Hirata. Notes on lax ends.arXiv:2210.01522,

Pith/arXiv arXiv
[2017]

[Man25] Diego Manco

Un- published manuscript. [Man25] Diego Manco. Coherence for pseudo commutative monads.arXiv:2509.15101,

arXiv
[2024]

Straightening for lax transformations and adjunctions of (∞,2)-categories.arXiv:2404.03971,

[AGH24] Fernando Abell´ an, Andrea Gagna, and Rune Haugseng. Straightening for lax transformations and adjunctions of (∞,2)-categories.arXiv:2404.03971,

Pith/arXiv arXiv
[2025]

Algebraic theories and (∞,1)-categories.arXiv:1011.3243,

[Cra10] James Cranch. Algebraic theories and (∞,1)-categories.arXiv:1011.3243,

Pith/arXiv arXiv
[2026]

Magmal characterisations of cocartesian categories.arXiv:2508.11615,

[Ark25] Nathanael Arkor. Magmal characterisations of cocartesian categories.arXiv:2508.11615,

Pith/arXiv arXiv

[1] [1]

Enhanced 2-categorical structures, two- dimensional limit sketches and the symmetry of internalisation.arXiv:2412.07475,

[ABK24] Nathanael Arkor, John Bourke, and Joanna Ko. Enhanced 2-categorical structures, two- dimensional limit sketches and the symmetry of internalisation.arXiv:2412.07475,

Pith/arXiv arXiv

[2] [3]

Free fibrations, lax colimits and kan extensions (∞,2)-categories.arXiv:2602.07604,

[AHM26] Fernando Abell´ an, Rune Haugseng, and Louis Martini. Free fibrations, lax colimits and kan extensions (∞,2)-categories.arXiv:2602.07604,

arXiv

[3] [1999]

Tensor product for symmetric monoidal categories.arXiv:0711.0324,

[Sch07] Vincent Schmitt. Tensor product for symmetric monoidal categories.arXiv:0711.0324,

Pith/arXiv arXiv

[4] [2010]

Nuclei of categories with tensor products.arXiv:0708.2761,

[Dav07] Alexei Davydov. Nuclei of categories with tensor products.arXiv:0708.2761,

Pith/arXiv arXiv

[5] [2012]

Linearly distributive fox theorem.arXiv:2506.02180,

64 [KB25] Rose Kudzman-Blais. Linearly distributive fox theorem.arXiv:2506.02180,

arXiv

[6] [2013]

Notes on lax ends.arXiv:2210.01522,

[Hir22] Kengo Hirata. Notes on lax ends.arXiv:2210.01522,

Pith/arXiv arXiv

[7] [2017]

[Man25] Diego Manco

Un- published manuscript. [Man25] Diego Manco. Coherence for pseudo commutative monads.arXiv:2509.15101,

arXiv

[8] [2024]

Straightening for lax transformations and adjunctions of (∞,2)-categories.arXiv:2404.03971,

[AGH24] Fernando Abell´ an, Andrea Gagna, and Rune Haugseng. Straightening for lax transformations and adjunctions of (∞,2)-categories.arXiv:2404.03971,

Pith/arXiv arXiv

[9] [2025]

Algebraic theories and (∞,1)-categories.arXiv:1011.3243,

[Cra10] James Cranch. Algebraic theories and (∞,1)-categories.arXiv:1011.3243,

Pith/arXiv arXiv

[10] [2026]

Magmal characterisations of cocartesian categories.arXiv:2508.11615,

[Ark25] Nathanael Arkor. Magmal characterisations of cocartesian categories.arXiv:2508.11615,

Pith/arXiv arXiv