Pith. sign in

REVIEW 3 major objections 5 minor 60 references

Layered monoidal theories embed several levels of abstraction — molecules, bits, qubits — into a single string diagram, with functor boxes and coboxes as derived notions rather than primitives.

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

2026-08-02 21:31 UTC pith:B2XJVPF3

load-bearing objection Layered monoidal theories is a genuine syntactic contribution, but the central semantic claims are deferred to a sequel and the key counit-section assumption is load-bearing. the 3 major comments →

arxiv 2602.19776 v2 pith:B2XJVPF3 submitted 2026-02-23 cs.LO

Layered Monoidal Theories I: Diagrammatic Algebra and Applications

Leo Lobski , Fabio Zanasi This is my paper
classification cs.LO MSC 18M0518N10
keywords layered monoidal theoriesstring diagramsmonoidal functorsfunctor boxesfunctor coboxesdeflational theorieslevels of abstraction2-categories
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.

This paper introduces layered monoidal theories, a syntactic framework in which several monoidal theories — each describing a system at a different level of abstraction — and the translations between them coexist within a single string diagram. It shows that the familiar 'functor box' notation for monoidal functors is not a primitive but a derived construction, and introduces its dual, the 'functor cobox'. Three classes of theories are defined — opfibrational, fibrational, and deflational — with deflational theories providing the 2-categorical structure that makes boxes and coboxes work. Six case studies (digital circuits, electrical circuits, quantum circuit extraction, concurrent processes, chemical reactions, and probability) demonstrate the formalism in action. If the framework holds, it gives engineers and scientists a uniform, precisely specified way to reason across coarser and finer descriptions of the same system.

Core claim

The paper's central claim is that a layered monoidal theory — a set of layers, generators between layers, and a monoidal signature per layer, extended by a recursive sorting procedure and 0-, 1-, and 2-equations — provides a uniform syntax in which multiple monoidal theories and functorial translations coexist inside one string diagram. In the two-layer, one-translation special case, the formalism recovers standard monoidal functorial semantics, with faithfulness of the translation equivalent to a single equation in the layered theory. Within deflational theories, functor boxes and their dual coboxes are derived as composites of refinement and coarsening maps, so the standard box equations b

What carries the argument

The central mechanism is the layered signature together with its three-level term calculus: types (0-cells), terms (1-cells), and 2-terms, each with their own equations. The load-bearing structure is the deflational theory, which glues an opfibrational and a fibrational theory along the internal terms and equips every coarsening map with an adjoint refinement via structural 2-cells (unit eta and counit epsilon). Functor boundaries (drawn as ◀f and ▷f) are the 1-cells that carry terms between layers; a functor box is shown to be a window ◀f·x·▷f through which the internal term x slides, using the structural adjunction 2-cells. Coboxes reverse the order of the same two maps, giving a term that

Load-bearing premise

The central claim depends on an assumption the paper states but does not prove here: that every 'zooming refinement' step in a deflational theory can be reversed (has a section), with the justification deferred to a companion paper; if such reversals do not always exist, the functor-box and cobox derivations lose their foundation.

What would settle it

Construct (or extract from the sequel) a deflational theory in which some counit 2-cell provably has no section, and test whether the bidirectional 2-cells of Proposition 4.2 still hold there. If they fail, the box/cobox decomposition genuinely depends on the assumption; if they hold, the assumption can be dropped from the framework's statement. Alternatively, exhibiting a 1-cell equal in Tε (the theory with sections added) but not in T would show the assumption changes the equational theory of layered terms.

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

Share X Bluesky LinkedIn Reddit HN

If this is right

  • The two-layer, one-translation case of a layered theory reproduces standard monoidal functorial semantics, and faithfulness of the induced functor becomes equivalent to a single equation (4.12) in the layered theory.
  • Functor boxes become derived notions: they decompose into coarsening, internal term, and refinement, and the usual functor equations are provable instead of imposed.
  • A dual functor cobox emerges, allowing a diagram to inspect a sub-diagram's finer semantics without translating the whole picture; a cobox also detects whether the functor it represents is faithful.
  • Each of the six case studies — arithmetic circuits, impedance boxes in electrical circuits, MBQC-to-circuit extraction, CCS reduction vs LTS semantics, glucose phosphorylation, and probabilistic conditionals — is expressible as a layered theory, exposing multi-level structure in systems previously treated as flat.
  • The framework is neutral between reductionism and emergentism: it formalises movement between abstraction levels without committing to which level is fundamental.

Where Pith is reading between the lines

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

  • If the framework is right, workflows that currently switch informal notation when changing resolution — from a circuit diagram to its gate-level implementation, or from reaction schemes to molecular graphs — could share one syntax, turning 'zoom in' and 'zoom out' into formal diagrammatic operations.
  • The cobox's faithfulness-detection property suggests a general proof technique: to prove a translation faithful, add the single equation (4.12) to the layered theory and reason there; this could mechanise completeness arguments across the six domains.
  • The infinite-layer construction for simple arithmetic circuits (one layer per wire width n) hints that other hierarchical parameterisations — precision levels, time scales, or finite approximations — could be indexed as layers, making nested approximations a single diagram.
  • The paper's own discussion notes the formalism is best suited to discrete systems with few parts; whether continuum coarse-graining (for instance statistical mechanics versus thermodynamics) can be captured remains an open stress test for layered theories.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper introduces layered monoidal theories (Definition 3.18) as a multi-level extension of monoidal theories: a layered signature indexes monoidal signatures by a set of layers, with generators between layers, and the theory is equipped with 0-, 1- and 2-equations. Three classes are singled out—opfibrational, fibrational, and deflational theories (Subsection 3.4)—the last adding structural 2-cells that make the fibrational counterparts of the opfibrational generators right adjoints. The core syntactic claim is that in deflational theories, the usual functor boxes decompose as 'cowindows' (Proposition 4.2), and the dual notion of cobox can be defined (Definition 4.6). Six case studies are presented: digital circuits, electrical circuits, ZX-calculus/MBQC, CCS, chemical reactions, and probabilistic channels. The semantic theory is explicitly deferred to a companion paper [47]; the CCS soundness/completeness result is proved in Appendix A.

Significance. If the deferred semantic facts hold, the framework is a genuinely useful unification: it gives a precise syntax for reasoning with several monoidal theories and translations in one diagram, recovers functor boxes and coboxes as derived notions, and provides a common lens for a wide range of existing string-diagrammatic formalisms. The definitions are rigorous and several structural results are proved in the text, and the CCS appendix contains a real soundness/completeness argument. The main value is conditional, however: the box/cobox machinery and several applications rely on the unproved counit-section assumption and on the existence of the intended semantic models, both of which are deferred to [47].

major comments (3)
  1. [Section 3.4, after Definition 3.33] The paper assumes that every counit 2-cell in a deflational theory has a section κ, with the statement that T and Tε have the same models deferred to [47]. This assumption is load-bearing: Proposition 4.2 uses κ to construct the bidirectional 2-cell between a cowindow and a functor box, and Corollaries 4.3–4.4 and the cobox equation in Definition 4.6 inherit that dependence. Adjoining a section to a counit is additional 2-cell structure and need not preserve the 2-categorical model theory. No model check is given here. The central claim that functor boxes are derived notions is therefore conditional on a nontrivial semantic fact. Please either prove the needed statement (or a sufficient special case), or restate Section 4 with the precise hypotheses under which each result holds.
  2. [Section 5.5, glucose phosphorylation] The derivation in Figure 20 is constructed so that restrictions (1) and (2) hold, but those restrictions are not part of the layered theory—the text explicitly says they are properties of the chosen term, and that without them the reaction is derivable whenever atoms and charge coincide. As stated, the example does not demonstrate that the layered theory formalizes reaction mechanisms; it shows only that a specially decorated term can be chosen. The restrictions need to be built into the theory (e.g., by restricting the generators or by defining a subtheory), or the claim about the example must be weakened accordingly.
  3. [Section 5.6 and Definition 3.20] The conditional box B is not a functor, and the paper says the functoriality equations must be dropped. But the deflational theories of Section 3 are defined by the structural equations, including functoriality of boxes; hence this example is not an instance of the framework studied in Section 4, and the diagrammatic reasoning with the conditional box is not covered by the paper's main results. The text asserts that the terms may still be used, but no soundness theorem is given for layered theories with non-functorial generators. Please state the exact class of layered theories that admits such generators and prove the relevant soundness, or recast the conditional box as a separate, precisely defined extension.
minor comments (5)
  1. [Section 4, introduction] Typo: 'peak' should be 'peek'.
  2. [Appendix A, first sentence] The phrase 'whore arity and coarity' should read 'whose arity and coarity'.
  3. [Section 2.1] Typo: 'Appandix' should be 'Appendix'.
  4. [Throughout] Several equation references are made to labels that are not displayed in the text (e.g., Equation (4.12) in the discussion before Proposition 4.11). Please add the displayed equation or adjust the referencing.
  5. [Section 5.5, Figure 20] The notation in the derivation, such as '𝐸46 𝐶12...' and the subscripted variables, is hard to verify without a detailed explanation. Please clarify the correspondence between the derivation and the generators of Disc and Scheme, or expand the figure caption.

Circularity Check

1 steps flagged

The syntactic framework is independently defined and the CCS theorem is proved, but the Section 4 box/cobox development rests on a deferred, same-author assumption about counit sections, making the advertised 'derived functor boxes' conditional rather than self-contained.

specific steps
  1. self citation load bearing [Section 3.4, after Definition 3.33; used in Proposition 4.2]
    "While T and Tε are, in general, not equivalent as deflational theories, in [47] we will show that they have the same (op)indexed monoidal categories as models. Since these are the main models of interest, in Sections 4 and 5 we will, therefore, make the assumption that the counits in Figure 10 have sections κ, as this will allow us to derive more properties within the formalism of deflational theories."

    Proposition 4.2 proves the key bidirectional 2-cell by applying the counit ε, 'which has a section κ'. The existence of κ is not derived or checked here: it is an assumption, semantically deferred to [47], a companion paper by the same authors. Corollaries 4.3/4.4 and Definition 4.6 inherit this assumption, and the Section 5 applications then use boxes/coboxes. Hence the advertised conclusion that functor boxes are 'a derived notion rather than a primitive one' rests on a self-cited, unverified premise: without κ the decomposition collapses. The paper is transparent about the deferral, so this is load-bearing self-citation rather than an equation-level circularity.

full rationale

No equation in the paper is derived from itself by construction. The CCS soundness/completeness results are proved by induction in Appendix A, the probabilistic conditional construction verifies the defining universal property, and the two-layer recovery of monoidal functorial semantics is a genuine theorem. The main circularity-adjacent issue is the counit-section assumption of Section 3.4: it is load-bearing for Proposition 4.2 and the downstream cobox equations, and its semantic justification is deferred to the authors' own companion paper [47]. The chemistry example is also explicitly reverse-engineered: restrictions (1) and (2) are chosen so that the intended phosphorylation reaction is derivable, and the paper admits that without them the reaction is always derivable from conservation laws. This is a presentational circularity rather than a predictive claim. Overall, the framework has substantial independent syntactic content, so the paper does not reduce to its inputs; the score reflects the genuine dependence on an unverified, self-cited semantic assumption.

Axiom & Free-Parameter Ledger

0 free parameters · 5 axioms · 3 invented entities

The paper contributes a new syntactic framework rather than empirically fitted parameters. The main 'free choices' are structural: the assumed counit sections, the imported theories, and the ad hoc restrictions in the chemistry example, all listed above. There are no numeric parameters fitted to data.

axioms (5)
  • ad hoc to paper Counit 2-cells in deflational theories have sections (κ).
    Section 3.4 explicitly assumes this, with proof deferred to [47]; Proposition 4.2 and related box/cobox results rely on it.
  • ad hoc to paper Semantic models of layered theories exist as (op)indexed monoidal categories.
    The paper states it 'contains no explicit discussion of semantic models – this is the topic of Part II [47]'.
  • domain assumption Imported monoidal theories (ZX-calculus, graphical affine algebra, FinStoch, CCS) are sound as given.
    The case studies take existing theories and their equations from the literature without re-proving them.
  • domain assumption Chemical disconnection rules capture all charge- and matter-preserving rearrangements.
    Relied on in Section 5.5, credited to the authors' own [25].
  • ad hoc to paper The conditional box B may be used in a layered signature even though it is not a monoidal functor.
    Section 5.6: 'Note that B is not a functor – hence we cannot assume the functoriality equations of a layered theory', meaning the example does not fit the general theory without modification.
invented entities (3)
  • Layered monoidal theory no independent evidence
    purpose: A formal syntax mixing several monoidal theories and translations in one string diagram.
    The central new formalism; no external falsifiable handle.
  • Functor cobox no independent evidence
    purpose: Dual of a functor box; acts on the exterior of a diagram, used to 'peek' into semantics.
    Definition 4.6; an internal mathematical notion.
  • Deflational structural 2-cells (unit/counit) no independent evidence
    purpose: Express refinements as adjunctions between opfibrational and fibrational terms.
    Definition 3.32; underlies the box/cobox calculus.

reviewed 2026-08-02 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Layered Monoidal Theories I: Diagrammatic Algebra and Applications." pith.science (2026). https://pith.science/paper/B2XJVPF3

@misc{pith2026260219776,
  author       = {Pith},
  title        = {Pith review of: Layered Monoidal Theories I: Diagrammatic Algebra and Applications},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/B2XJVPF3}},
  note         = {Machine review of arXiv:2602.19776}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

We develop layered monoidal theories -- a generalisation of monoidal theories combining formal descriptions of a system at different levels of abstraction. Via their representation as string diagrams, monoidal theories provide a graphical formalism to reason algebraically about information flow in models across different fields of science. Layered monoidal theories allow mixing several monoidal theories (together with translations between them) within the same string diagram, while retaining mathematical precision and semantic interpretability. We develop the mathematical foundations of layered monoidal theories, as well as providing several instances of our approach, including digital and electrical circuits, quantum processes, chemical reactions, concurrent processes, and probability theory.

Figures

Figures reproduced from arXiv: 2602.19776 by Fabio Zanasi, Leo Lobski.

Figure 1
Figure 1. Figure 1: Left: an informal translation from a coarser language (English names) to a finer language (molecular graphs). Right: A [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Various features of layered monoidal theories. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Rules for generating the basic terms of a layered signature [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: 1-equations defining monoidal functors inside the fibres [PITH_FULL_IMAGE:figures/full_fig_p012_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: The structural 2-equations. Here id𝑇 and id𝑆 are the identity terms on the types 𝑇 and 𝑆, while id𝜀 is the identity term on 𝜀 : 𝜀 obtained by the rule (ext-unit). two free models. The functor is free in the sense that all the terms obtained by applying 𝑓 to the internal terms with type 𝜔 (the (int-box) rule) remain uninterpreted, and are simply added as generators to the monoidal category defined by the in… view at source ↗
Figure 6
Figure 6. Figure 6: Rules for generating the opfibrational terms [PITH_FULL_IMAGE:figures/full_fig_p014_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: 1-equations defining monoidal categories [PITH_FULL_IMAGE:figures/full_fig_p015_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: 1-equations defining monoidal functors 𝐴 : 𝜔 𝑓 ∈ Ω(𝜔, 𝜏 ) (ext-gen-op) ▷𝑓 : (𝑓 (𝐴) : 𝜏 | 𝐴 : 𝜔) 𝜔 ∈ Ω (counit) : (𝜀 : 𝜔 | 𝜀 : 𝜀 ) 𝐴 : 𝜔 (codiag-unit) : (𝜀 : 𝜀 | 𝐴 : 𝜔) 𝐴 : 𝜔 𝐵 : 𝜔 (comonoid) : (𝐴𝐵 : 𝜔 | 𝐴 : 𝜔, 𝐵 : 𝜔) 𝐴 : 𝜔 (codiag) : (𝐴 : 𝜔, 𝐴 : 𝜔 | 𝐴 : 𝜔) [PITH_FULL_IMAGE:figures/full_fig_p015_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Rules for generating the fibrational terms [PITH_FULL_IMAGE:figures/full_fig_p015_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Structural 2-cells for deflational theories [PITH_FULL_IMAGE:figures/full_fig_p016_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Layered theory defining the simple arithmetic circuits. [PITH_FULL_IMAGE:figures/full_fig_p022_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: A simple arithmetic logic unit. 5.2 Electrical circuits We formalise the notion of an impedance box, introduced in the study of compositional electrical circuit theory by Boisseau & Sobociński [7] and Boisseau [6]. We find that it corresponds to a composition of a box with a cobox in a layered monoidal theory generated by translating electrical circuits to graphical affine algebra (Definition 5.7). Within… view at source ↗
Figure 13
Figure 13. Figure 13: Equations between terms defining the functor [PITH_FULL_IMAGE:figures/full_fig_p024_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Equations between terms defining the functor [PITH_FULL_IMAGE:figures/full_fig_p025_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Generating 2-cells and equations in MBQC+LC. (5) to each vertex 𝑢 ∈ 𝑉𝐺 \𝑂𝐺 with 𝜆𝐺 (𝑃, 𝛼), attach the measurement angle 𝛼 in the plane 𝑃, as specified in [PITH_FULL_IMAGE:figures/full_fig_p031_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: The structural generators of Red. The generators of Red are given by: (1) the gener￾ators (5.29), where 𝑎 ∈ 𝐴, capturing the produc￾tion rules s, r and p of Definition 5.26, and (2) the structural generators, consisting of the symmetry and the generators in [PITH_FULL_IMAGE:figures/full_fig_p034_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Layered theory for CCS. The equations in the first four rows hold for any structural generators in [PITH_FULL_IMAGE:figures/full_fig_p035_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: The four disconnection rules. We denote the vertex names by either positive integers or lowercase Latin letters, as appropriate to the situation. While formally the set of vertex labels can contain arbitrary symbols, in what follows we shall assume that At contains a symbol for each main-group element of the periodic table: {𝐻,𝐶,𝑂, 𝑃, . . . } ⊆ At. For this reason, we will also refer to At as the atom lab… view at source ↗
Figure 19
Figure 19. Figure 19: The functor Name → Disc. We adopt the usual organic chemistry convention that an unlabelled vertex stands for a carbon atom with an appropriate number of hydrogen atoms attached to make the valence add up to 4. For clarity, we omit the vertex names, as their precise choice is immaterial: the only requirement we impose is that the vertex sets of all five graphs are pairwise disjoint. We are now ready to sh… view at source ↗
Figure 20
Figure 20. Figure 20: Glucose phosphorylation as a layered term. [PITH_FULL_IMAGE:figures/full_fig_p041_20.png] view at source ↗
Figure 21
Figure 21. Figure 21: Equations of the coparametric monoidal theory. [PITH_FULL_IMAGE:figures/full_fig_p052_21.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

60 extracted references · 13 canonical work pages · 2 internal anchors

  1. [1]

    Andersen, Christoph Flamm, Daniel Merkle, and Peter F

    Jakob L. Andersen, Christoph Flamm, Daniel Merkle, and Peter F. Stadler. 2019. Chemical Transformation Motifs – Modelling Pathways as Integer Hyperflows.IEEE/ACM transactions on computational biology and bioinformatics16, 2 (2019), 510–523. doi:10.1109/TCBB.2017.2781724

    arXiv 2019

  2. [2]

    Miriam Backens. 2014. The ZX-calculus is complete for stabilizer quantum mechanics.New Journal of Physics16, 9 (2014). doi:10.1088/1367- 2630/16/9/093021

    doi:10.1088/1367- 2014

  3. [3]

    Backens, H

    M. Backens, H. Miller-Bakewell, G. De Felice, L. Lobski, and J. van de Wetering. 2021. There and back again: A circuit extraction tale.Quantum5, 421 (2021). doi:10.22331/q-2021-03-25-421

    doi:10.22331/q-2021-03-25-421 2021

  4. [4]

    A Compositional Framework for Passive Linear Networks

    John C. Baez and Brendan Fong. 2015. A Compositional Framework for Passive Linear Networks. doi:10.48550/ARXIV.1504.05625

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1504.05625 2015

  5. [5]

    Douglas, Christopher J

    Bruce Bartlett, Christopher L. Douglas, Christopher J. Schommer-Pries, and Jamie Vicary. 2015. Modular categories as representations of the 3-dimensional bordism 2-category. doi:10.48550/ARXIV.1509.06811

    Pith/arXiv arXiv doi:10.48550/arxiv.1509.06811 2015

  6. [6]

    2023.Graphical Electrical Circuit Theory

    Guillaume Boisseau. 2023.Graphical Electrical Circuit Theory. Ph. D. Dissertation. Department of Computer Science. https://www.proquest.com/ dissertations-theses/graphical-electrical-circuit-theory/docview/3165317412/se-2

    arXiv 2023

  7. [7]

    Guillaume Boisseau and Paweł Sobociński. 2021. String Diagrammatic Electrical Circuit Theory. doi:10.48550/ARXIV.2106.07763

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2106.07763 2021

  8. [8]

    Filippo Bonchi and Cipriano Junior Cioffo. 2026. Tapes as Stochastic Matrices of String Diagrams. arXiv:2601.01472

    arXiv 2026

  9. [9]

    Filippo Bonchi, Robin Piedeleu, Pawel Sobociński, and Fabio Zanasi. 2019. Graphical Affine Algebra. In2019 34th Annual ACM/IEEE Symposium on Logic in Computer Science (LICS). IEEE, 1–12. doi:10.1109/LICS.2019.8785877

    arXiv 2019

  10. [10]

    2021.A Survey of Compositional Signal Flow Theory

    Filippo Bonchi, Paweł Sobociński, and Fabio Zanasi. 2021.A Survey of Compositional Signal Flow Theory. Springer International Publishing, Cham, 29–56. doi:10.1007/978-3-030-81701-5_2

    doi:10.1007/978-3-030-81701-5_2 2021

  11. [11]

    Dylan Braithwaite and Mario Román. 2023. Collages of String Diagrams.Electronic Proceedings in Theoretical Computer Science397 (2023), 39–53. doi:10.4204/eptcs.397.3

    work page doi:10.4204/eptcs.397.3 2023

  12. [12]

    Daniel E Browne, Elham Kashefi, Mehdi Mhalla, and Simon Perdrix. 2007. Generalized flow and determinism in measurement-based quantum computation.New Journal of Physics9, 8 (2007). doi:10.1088/1367-2630/9/8/250

    doi:10.1088/1367-2630/9/8/250 2007

  13. [13]

    Bob Coecke, Mehrnoosh Sadrzadeh, and Stephen Clark. 2010. Mathematical Foundations for a Compositional Distributional Model of Meaning. arXiv:1003.4394

    Pith/arXiv arXiv 2010

  14. [14]

    Cole Comfort. 2024. Notes on profunctors and compact multiplicative linear logic. https://colecomfort.github.io/pdfs/notes.pdf

    2024

  15. [15]

    G. S. H. Cruttwell, Bruno Gavranović, Neil Ghani, Paul Wilson, and Fabio Zanasi. 2021. Categorical Foundations of Gradient-Based Learning. arXiv:2103.01931

    Pith/arXiv arXiv 2021

  16. [16]

    Vincent Danos, Elham Kashefi, and Prakash Panangaden. 2007. The measurement calculus.J. ACM54, 2 (2007). doi:10.1145/1219092.1219096

    arXiv 2007

  17. [17]

    Vincent Danos, Elham Kashefi, Prakash Panangaden, and Simon Perdrix. 2009. Extended Measurement Calculus. InSemantic Techniques in Quantum Computation, Simon Gay and Ian Mackie (Eds.). Cambridge University Press, 235–310. doi:10.1017/CBO9781139193313.008

    work page doi:10.1017/cbo9781139193313.008 2009

  18. [18]

    Dunn III

    Lawrence H. Dunn III. 2015.Profunctor Semantics for Linear Logic. Master’s thesis. Department of Computer Science. https://www.cs.ox.ac.uk/ people/aleks.kissinger/theses/bob/Dunn.pdf

    2015

  19. [19]

    Matt Earnshaw and Mario Román. 2024. Context-Free Languages of String Diagrams. arXiv:2404.10653

    Pith/arXiv arXiv 2024

  20. [20]

    Markus Englberger and Devendra Singh Dhami. 2025. Causal Abstractions, Categorically Unified. arXiv:2510.05033

    arXiv 2025

  21. [21]

    Spivak, and Rémy Tuyéras

    Brendan Fong, David I. Spivak, and Rémy Tuyéras. 2019. Backprop as Functor: A compositional perspective on supervised learning. doi:10.48550/ arXiv.1711.10455

    Pith/arXiv arXiv doi:10.48550/arxiv.1711.10455 2019

  22. [22]

    Tobias Fritz. 2009. A presentation of the category of stochastic matrices. arXiv:0902.2554 Manuscript submitted to ACM Layered Monoidal Theories I: Diagrammatic Algebra and Applications 47

    Pith/arXiv arXiv 2009

  23. [23]

    Tobias Fritz. 2020. A synthetic approach to Markov kernels, conditional independence and theorems on sufficient statistics.Advances in Mathematics 370 (2020), 107239. doi:10.1016/j.aim.2020.107239

    arXiv 2020

  24. [24]

    Tobias Fritz and Eigil Fjeldgren Rischel. 2020. Infinite products and zero-one laws in categorical probability.CompositionalityVolume 2 (2020) (2020). doi:10.32408/compositionality-2-3

    doi:10.32408/compositionality-2-3 2020

  25. [25]

    Ella Gale, Leo Lobski, and Fabio Zanasi. 2025. A categorical model for organic chemistry.Theoretical Computer Science1032 (2025), 115084. doi:10.1016/j.tcs.2025.115084

    arXiv 2025

  26. [26]

    Bruno Gavranović. 2019. Compositional Deep Learning. arXiv:1907.08292

    Pith/arXiv arXiv 2019

  27. [27]

    Gregor Goessler and Daniel Le Metayer. 2015. A general framework for blaming in component-based systems.Science of computer programming 113, Part 3 (2015), 223–235. doi:10.1016/j.scico.2015.06.010hal-01211484

    work page doi:10.1016/j.scico.2015.06.010hal-01211484 2015

  28. [28]

    Amar Hadzihasanovic, Kang Feng Ng, and Quanlong Wang. 2018. Two complete axiomatisations of pure-state qubit quantum computing. In Proceedings of the 33rd Annual ACM/IEEE Symposium on Logic in Computer Science (LICS ’18). Association for Computing Machinery, 502–511. doi:10.1145/3209108.3209128

    arXiv 2018

  29. [29]

    Hartwell, John J

    Leland H. Hartwell, John J. Hopfield, Stanislas Leibler, and Andrew W. Murray. 1999. From molecular to modular cell biology.Nature (London)402, 6761 (1999), C47–C52. doi:10.1038/35011540

    work page doi:10.1038/35011540 1999

  30. [30]

    James Hefford and Cole Comfort. 2023. Coend Optics for Quantum Combs.Electronic Proceedings in Theoretical Computer Science380 (2023), 63–76. doi:10.4204/eptcs.380.4

    work page doi:10.4204/eptcs.380.4 2023

  31. [31]

    2019.Categories for Quantum Theory: An Introduction

    Chris Heunen and Jamie Vicary. 2019.Categories for Quantum Theory: An Introduction. Oxford University Press. doi:10.1093/oso/9780198739623.001. 0001

    arXiv 2019

  32. [32]

    2023.Introducing String Diagrams: The Art of Category Theory

    Ralf Hinze and Dan Marsden. 2023.Introducing String Diagrams: The Art of Category Theory. Cambridge University Press. doi:10.1017/9781009317825

    work page doi:10.1017/9781009317825 2023

  33. [33]

    2019.External traced monoidal categories

    Nick Hu. 2019.External traced monoidal categories. Master’s thesis. Department of Computer Science. https://www.cs.ox.ac.uk/files/11696/project.pdf

    2019

  34. [34]

    Nick Hu and Jamie Vicary. 2021. Traced Monoidal Categories as Algebraic Structures in Prof.Electronic Proceedings in Theoretical Computer Science 351 (2021), 84–97. doi:10.4204/eptcs.351.6

    work page doi:10.4204/eptcs.351.6 2021

  35. [35]

    Bart Jacobs. 2025. Structured Probabilistic Reasoning. https://www.cs.ru.nl/B.Jacobs/PAPERS/ProbabilisticReasoning.pdf

    2025

  36. [36]

    Bart Jacobs, Aleks Kissinger, and Fabio Zanasi. 2021. Causal inference via string diagram surgery: A diagrammatic approach to interventions and counterfactuals.Math. Struct. Comput. Sci.31, 5 (2021), 553–574. doi:10.1017/S096012952100027X

    doi:10.1017/s096012952100027x 2021

  37. [37]

    Bart Jacobs, Márk Széles, and Dario Stein. 2025. Compositional Inference for Bayesian Networks and Causality.Electronic Notes in Theoretical Informatics and Computer ScienceVolume 5-Proceedings of... (2025). doi:10.46298/entics.17029

    work page doi:10.46298/entics.17029 2025

  38. [38]

    Bart Jacobs and Fabio Zanasi. 2020. The Logical Essentials of Bayesian Reasoning. InFoundations of Probabilistic Programming, Gilles Barthe, Joost-Pieter Katoen, and Alexandra Silva (Eds.). Cambridge University Press, 295–332. doi:10.1017/9781108770750.010

    doi:10.1017/9781108770750.010 2020

  39. [39]

    Emmanuel Jeandel, Simon Perdrix, and Renaud Vilmart. 2020. Completeness of the ZX-Calculus.Logical Methods in Computer Science16 (2020). Issue 2. doi:10.23638/LMCS-16(2:11)2020

    work page doi:10.23638/lmcs-16(2:11)2020 2020

  40. [40]

    George Kaye. 2025. Foundations of Digital Circuits: Denotation, Operational, and Algebraic Semantics. arXiv:2502.08497

    Pith/arXiv arXiv 2025

  41. [41]

    Aleks Kissinger and Sander Uijlen. 2019. A categorical semantics for causal structure.Logical Methods in Computer ScienceVolume 15, Issue 3 (Aug. 2019). doi:10.23638/lmcs-15(3:15)2019

    doi:10.23638/lmcs-15(3:15)2019 2019

  42. [42]

    Jean Krivine. 2017. Systems Biology.ACM SIGLOG News4, 3 (jul 2017), 43–61. doi:10.1145/3129173.3129182

    arXiv 2017

  43. [43]

    Jean Krivine. 2019. Physical systems, composite explanations and diagrams. https://youtu.be/w8vTTvA0HTE and https://www.cl.cam.ac.uk/events/ syco/strings3-syco5/

    2019

  44. [44]

    Jonathan Laurent, Jean Yang, and Walter Fontana. 2018. Counterfactual resimulation for causal analysis of rule-based models. InProceedings of the 27th International Joint Conference on Artificial Intelligence (IJCAI’18). AAAI Press, 1882–1890. doi:10.5555/3304889.3304920

    arXiv 2018

  45. [45]

    Gabriele Lobbia, Wojciech Różowski, Ralph Sarkis, and Fabio Zanasi. 2025. Quantitative Monoidal Algebra: Axiomatising Distance with String Diagrams. arXiv:2410.09229

    Pith/arXiv arXiv 2025

  46. [46]

    Leo Lobski and Fabio Zanasi. 2023. String Diagrams for Layered Explanations.Electronic Proceedings in Theoretical Computer Science380 (2023), 362–382. doi:10.4204/eptcs.380.21

    work page doi:10.4204/eptcs.380.21 2023

  47. [47]

    Leo Lobski and Fabio Zanasi. 2026. Layered Monoidal Theories II: Fibrational Semantics. To appear

    2026

  48. [48]

    Robin Lorenz and Sean Tull. 2023. Causal models in string diagrams. arXiv:2304.07638

    Pith/arXiv arXiv 2023

  49. [49]

    Robin Lorenz and Sean Tull. 2026. Causal and Compositional Abstraction. arXiv:2602.16612

    arXiv 2026

  50. [50]

    Paul-André Mellies. 2006. Functorial Boxes in String Diagrams. InComputer Science Logic (Lecture Notes in Computer Science, Vol. 4207). Springer Berlin Heidelberg, Berlin, Heidelberg, 1–30. doi:10.1007/11874683_1

    doi:10.1007/11874683_1 2006

  51. [51]

    1980.A Calculus of Communicating Systems

    Robin Milner. 1980.A Calculus of Communicating Systems. Lecture Notes in Computer Science, Vol. 92. Springer Berlin Heidelberg, Berlin, Heidelberg. doi:10.1007/3-540-10235-3

    doi:10.1007/3-540-10235-3 1980

  52. [52]

    Nashed, Saaduddin Mahmud, Claudia V

    Samer B. Nashed, Saaduddin Mahmud, Claudia V. Goldman, and Shlomo Zilberstein. 2023. Causal Explanations for Sequential Decision Making Under Uncertainty. arXiv:2205.15462

    Pith/arXiv arXiv 2023

  53. [53]

    2009.The Logic of Structure-Based Counterfactuals(2 ed.)

    Judea Pearl. 2009.The Logic of Structure-Based Counterfactuals(2 ed.). Cambridge University Press, 201–258. doi:10.1017/CBO9780511803161.009

    work page doi:10.1017/cbo9780511803161.009 2009

  54. [54]

    Robin Piedeleu, Mateo Torres-Ruiz, Alexandra Silva, and Fabio Zanasi. 2024. A Complete Axiomatisation of Equivalence for Discrete Probabilistic Programming. arXiv:2408.14701 Manuscript submitted to ACM 48 Leo Lobski and Fabio Zanasi

    Pith/arXiv arXiv 2024

  55. [55]

    2025.An Introduction to String Diagrams for Computer Scientists

    Robin Piedeleu and Fabio Zanasi. 2025.An Introduction to String Diagrams for Computer Scientists. Cambridge University Press. doi:10.1017/ 9781009625715

    2025

  56. [56]

    Mario Román. 2021. Open Diagrams via Coend Calculus.Electronic Proceedings in Theoretical Computer Science333 (feb 2021), 65–78. doi:10.4204/ eptcs.333.5

    2021

  57. [57]

    1991.Life itself : a comprehensive inquiry into the nature, origin, and fabrication of life

    Robert Rosen. 1991.Life itself : a comprehensive inquiry into the nature, origin, and fabrication of life. Columbia University Press, New York

    1991

  58. [58]

    Peter Selinger. 2010. A Survey of Graphical Languages for Monoidal Categories.Lecture Notes in Physics813 (2010). doi:10.1007/978-3-642-12821-9_4

    doi:10.1007/978-3-642-12821-9_4 2010

  59. [59]

    John van de Wetering. 2020. ZX-calculus for the working quantum computer scientist. arXiv:2012.13966

    Pith/arXiv arXiv 2020

  60. [60]

    deparemeterisation

    Fabio Zanasi. 2015.Interacting Hopf Algebras: the theory of linear systems. Ph. D. Dissertation. Laboratoire de l’Informatique du Parallélisme. doi:10.48550/ARXIV.1805.03032 A Proofs for Subsection 5.4 Here we give the proofs omitted in the discussion of the calculus of communicating systems. Lemma A.1 (Lemma 5.32).Let 𝑃 and𝑄 be processes. The following a...

    Pith/arXiv arXiv doi:10.48550/arxiv.1805.03032 2015

This paper was first reviewed by deepseek-v4-flash on August 2, 2026.