Pith. sign in

REVIEW 3 major objections 4 minor 19 references

Every closed dynamical system is a functor from an evolution shape to a state category, and Lyapunov stability and convergence reduce to filter conditions.

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 ↗

A categorical framework defines dynamical systems as functors from abstract evolution shapes to coefficient categories, with convergence and Lyapunov stability expressed through cosieve filters and sublevel neighbourhoods.

T0 review reviewed 2026-08-01 challenge →

load-bearing objection A clean categorical framework for dynamics whose Lyapunov convergence claim is weaker than advertised: the key basis condition is built into the definition, so classical recovery only holds under extra assumptions. the 3 major comments →

arxiv 2607.17455 v1 pith:UZPIMWCJ submitted 2026-07-20 math.CT math.DS

Dynamical Systems as Functorial Realisations of Abstract Evolution Shapes

Bangxin Wang This is my paper
classification math.CT math.DS MSC 18A0518A2518B99
keywords categorical dynamical systemsfunctorsevolution shapeseventuality filtersneighbourhood filtersLyapunov stabilityLyapunov decayguarded hybrid systems
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 reading

This paper argues that the syntax of a dynamical system—which evolutions are admissible—can be separated from its semantics—what its state spaces are—by modelling a closed system as a functor X:S→C from a small category S (the evolution shape) to a coefficient category C. In this functorial setting, invariant subsystems, equilibria, and orbits are defined categorically, and convergence is expressed using two exogenous filter structures: an eventuality filter on S (which evolutions count as 'eventually') and a neighbourhood filter of the invariant subsystem (what counts as 'close'). The paper's categorical Lyapunov principle is that a Lyapunov function—a measurement vanishing on the invariant subsystem and non-increasing along evolutions—produces forward-invariant sublevel neighbourhoods; if these sublevels form a basis of the neighbourhood filter, Lyapunov stability follows, and Lyapunov decay implies convergence. The payoff is a single framework covering autonomous, non-autonomous, switched, hybrid, and stochastic systems, recovering the classical Lyapunov method in standard discrete-time and guarded-hybrid examples.

Core claim

The discovery is that Lyapunov theory's essence is not the numerical value of a Lyapunov function but the family of sublevel neighbourhoods it defines and their forward invariance under the dynamics. The paper formalises this as a Lyapunov basis: a basis of the neighbourhood filter of an invariant subsystem consisting of neighbourhoods that the system's evolutions carry into corresponding neighbourhoods at later stages. Lemma 4.4 shows such a basis implies Lyapunov stability. Then a categorical Lyapunov function—a compatible family of measurements L_s:X(s)→R that vanish on the invariant subsystem and are non-increasing along evolutions—is shown (Theorem 4.9) to produce such a basis via pullb

What carries the argument

The central object is the functor X:S→C, with S an abstract evolution shape (a small category whose objects are stages/modes and whose morphisms are admissible evolutions) and C a coefficient category of state spaces. Around it, two filter structures do the work: an eventuality filter E_s on each set of outgoing morphisms (a filter of cosieves on S, specifying which evolutions count as 'sufficiently far along'), and a neighbourhood filter N_X(I) of an invariant subsystem I (stagewise monomorphisms through which I factors). The Lyapunov machinery then sits on top: a Lyapunov measurement datum (R, 0, sublevel monomorphisms R_λ→R, preorder) and a categorical Lyapunov function L={L_s:X(s)→R} tha

Load-bearing premise

The convergence theorem assumes the Lyapunov function's sublevel neighbourhoods form a basis of the chosen neighbourhood filter of the invariant subsystem; if the filter is finer than what the sublevels can see, decay can hold without true convergence.

What would settle it

Take M=[0,1], f(x)=x/2, equilibrium x∞=0, Lyapunov L(x)=x. Use the standard eventuality filter (n≥N) but take the neighbourhood filter of 0 to be the discrete filter consisting of all subsets of [0,1] containing 0. Then L is a categorical Lyapunov function and every trajectory has Lyapunov decay (L(f^n(x)) = x/2^n → 0, so it eventually enters every [0,ε)), yet the trajectory never enters the neighbourhood {0}, so convergence as defined in the paper fails. This shows the basis condition (Definition 4.7(iii)) is required for Theorem 4.11.

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

If this is right

  • Every closed dynamical system—autonomous, non-autonomous, switched, hybrid, stochastic—can be written as X:S→C, making the admissible-evolution pattern an explicit, composable component of the model.
  • Lyapunov stability is equivalent to the existence of a forward-invariant basis of neighbourhoods; any categorical Lyapunov function supplies one, so stability follows from existence of such a function.
  • Lyapunov decay—eventual entry into every sublevel neighbourhood—implies convergence to the invariant subsystem whenever the sublevels form a basis of the chosen neighbourhood filter.
  • In the standard discrete-time and guarded-hybrid examples, these categorical criteria reduce to the classical Lyapunov conditions: a non-increasing measurement whose value tends to zero along trajectories.

Where Pith is reading between the lines

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

  • The syntax–semantics split suggests a compositional route: because the evolution shape and the coefficient category are independent, gluing systems could be done at the level of shapes (e.g., pushouts or spans of categories) without reworking the Lyapunov machinery—a direction the paper does not pursue.
  • The eventuality filter is exogenous, so the same functor can converge or not depending on the choice of 'eventually'; convergence claims in this framework are therefore always relative to an explicit filter, which is a feature for comparing different asymptotic notions (almost sure, in probability, etc.) in one language.
  • The convergence theorem is purely filter-theoretic and does not require metrics or topologies on state spaces, suggesting that Lyapunov theory could be developed in coefficient categories like Stoch, where sublevels might be defined by stochastic dominance rather than pointwise inequality.
  • If the sublevel-basis condition fails, Lyapunov decay no longer guarantees convergence even in Euclidean examples; a natural testable extension is to characterise which neighbourhood filters admit a Lyapunov basis, or to weaken the convergence conclusion to 'convergence in the topology generated by the Lyapunov measurement'.
Share X Bluesky LinkedIn Reddit HN

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 / 4 minor

Summary. The paper develops a categorical framework in which a closed dynamical system is a functor X:S→C from a small category S of admissible evolutions to a coefficient category C. It defines invariant subsystems, equilibria, and orbits in functorial terms, and formalizes convergence using an eventuality filter on S and a neighbourhood filter on the invariant subsystem. It then introduces Lyapunov measurement data and categorical Lyapunov functions (Definition 4.7), proves a stability theorem (Theorem 4.9) and a Lyapunov convergence criterion (Theorem 4.11), and claims that these recover the classical Lyapunov method in autonomous discrete-time and guarded hybrid examples.

Significance. The syntax–semantics separation is a clean organizing principle and the framework does encompass monoid, poset, switched, hybrid, and stochastic systems. The categorical statements are short and, under their explicit hypotheses, correctly proved. The paper would be a useful unifying presentation if the advertised recovery of classical Lyapunov theory were actually established. The main missing piece is the verification of Definition 4.7(iii) for the standard topological/semantic filter; the examples choose the filter to be generated by the Lyapunov sublevels themselves, which makes that condition true by fiat but does not imply topological convergence. With added properness assumptions or an explicit identification of the sublevel filter with the semantic filter, the framework would be sound and significant.

major comments (3)
  1. [§4.3, Example 4.12 (around Eq. (4.73))] The sublevel sets U^ε = {L<ε} are not generally a neighbourhood basis of the equilibrium. For M=R, L(x)=|x|/(1+|x|)^2 is continuous, positive definite, and non-increasing along f(x)=x for x≤1, f(x)=2x−1 for x≥1; 0 is a fixed point, but for x0=2 one has L(f^n(x0))→0 while f^n(x0)→∞. The sets U^ε are unbounded and do not form a basis of the usual topological filter at 0. Hence Lyapunov decay does not imply topological convergence. The claim that the categorical criterion 'recovers the classical Lyapunov condition' therefore requires an additional properness/bounded-sublevel assumption or a proof that the sublevel filter is a basis for the semantic filter. The same caveat applies to the stability statement in Theorem 4.9.
  2. [Definition 4.7(iii) and Theorem 4.11] When N_X(I) is taken to be the filter generated by the sublevels, as in Examples 4.12 and 4.13, condition (iii) holds by construction and Theorem 4.11 reduces to a restatement of Definition 4.10. The theorem's substantive content is entirely the basis condition. The paper should make this explicit and verify the condition for the standard topological or semantic neighbourhood filter; otherwise 'convergence' is only convergence in the sublevel filter.
  3. [Example 4.13] The same issue appears in the hybrid example. The neighbourhood filter N_X(I) is defined as the stagewise filter generated by the sublevel neighbourhoods U^ε_{q,t}, making the categorical Lyapunov criterion filter-relative. No argument shows that this filter coincides with, or is a basis for, the natural topological/hybrid neighbourhood filter of the invariant subsystem. Thus the claimed recovery of classical convergence for guarded hybrid systems is not established.
minor comments (4)
  1. [Section 3 title] 'Inv ariant subsystems' contains a typo; should be 'Invariant subsystems'.
  2. [§4.2, first paragraph] 'explain how thy produce' should read 'explain how they produce'.
  3. [References] The references [Ru00], [SSV20], and [ST17] are listed but do not appear to be cited in the text.
  4. [Example 4.12, final paragraph] The phrase 'the trajectory eventually enters every sublevel neighbourhood of the equilibrium' should be qualified as 'with respect to the chosen sublevel filter' unless the basis condition for the topological filter is proved.

Circularity Check

3 steps flagged

Categorical Lyapunov principle is definitional: the basis condition is assumed in Definition 4.7(iii), so Theorem 4.11 and the example 'recoveries' restate the convergence definition.

specific steps
  1. self definitional [Section 4.3, Theorem 4.11 (proof), with Definition 4.7(iii) and Definition 3.17]
    "Since the sublevel neighbourhoods of L form a basis for N_X(I), there exists λ∈Λ such that U^λ ≤ U. ... By the universal property of pullback, X(α)∘a factors through U^λ_t,→X(t). Since U^λ ≤ U, it follows that X(α)∘a factors through U_t,→X(t) ... This is precisely the definition of convergence of a to I."

    Definition 4.7(iii) already requires the sublevel stagewise neighbourhoods to be a basis of the chosen neighbourhood filter N_X(I). Definition 4.10 defines Lyapunov decay as eventual factoring through every sublevel R^λ, and Definition 3.17 defines convergence as eventual factoring through every U∈N_X(I). With the basis condition assumed, these are the same eventuality statement, connected only by the pullback that defines U^λ. The proof itself says the conclusion is 'precisely the definition of convergence'. Thus the central Lyapunov convergence criterion is an unpacking of the defining basis condition rather than a substantive dynamical result.

  2. self definitional [Section 4.2, Theorem 4.9 (proof) and Definition 4.7(iii)]
    "By definition of a categorical Lyapunov function, the sublevel neighbourhoods U^λ form a basis for N_X(I). It remains to show that each U^λ is forward-invariant."

    The basis part of the claimed Lyapunov-basis result is exactly condition 4.7(iii), i.e. it is assumed in the definition of a categorical Lyapunov function, not derived. The only new argument is forward-invariance, which follows immediately from non-increasingness and downward-closedness of the sublevels. Stability then follows by Lemma 4.4, which is a one-step unpacking of Definition 4.1. Hence the advertised 'categorical Lyapunov principle' for stability is largely a restatement of the definitions.

  3. other [Section 4.3, Example 4.12 (Autonomous discrete-time systems)]
    "Therefore they form a Lyapunov basis for the neighbourhood filter generated by the sublevel sets of L. With respect to this neighbourhood filter, L is a categorical Lyapunov function for ι, and the categorical Lyapunov stability principle recovers the usual sublevel-set form of Lyapunov stability."

    Here the neighbourhood filter is chosen to be generated by the sublevel sets themselves, so condition 4.7(iii) holds by fiat. The paper then says that Lyapunov decay 'recovers the classical Lyapunov condition for convergence' and that the trajectory 'eventually enters every sublevel neighbourhood'. This is a recovery only relative to the manufactured sublevel filter. No proof is given that this filter equals the usual topological neighbourhood filter of the equilibrium. For non-proper Lyapunov functions, sublevels need not form a topological basis, so the claimed recovery of the classical method is not established; the example obtains its conclusion by choosing the filter so that the definition applies.

full rationale

The paper is self-contained and does not rely on a self-citation chain: the categorical Lyapunov references [AMT25, AMM25] are background only, and no load-bearing uniqueness theorem is imported from the author's prior work. The formal development is coherent and the proofs are valid relative to the definitions. The circularity is structural: Definition 4.7(iii) builds into 'categorical Lyapunov function' the requirement that sublevel stagewise neighbourhoods form a basis of the chosen neighbourhood filter. Theorem 4.9 then derives that these sublevels form a Lyapunov basis, but its proof begins by quoting that defining condition. Theorem 4.11 is even closer to a tautology: Lyapunov decay is defined as eventual factoring through every sublevel value, convergence is defined as eventual factoring through every neighbourhood, and the basis condition identifies these two collections; the proof's final sentence says the result is 'precisely the definition of convergence'. In the examples, the neighbourhood filter is chosen to be generated by the sublevels, so condition (iii) holds by construction; the paper does not prove that this filter is the ordinary topological neighbourhood filter. Indeed, for L(x)=|x|/(1+|x|)^2 on R, the sublevel {L<ε} is unbounded, so it is not contained in a small topological neighbourhood of 0; one can have L(f^n(x))→0 while f^n(x)→∞. Thus the claimed recovery of the classical Lyapunov method is not established beyond the sublevel filter. This is a partially definitional paper: the central 'prediction' reduces by construction, though the framework itself is logically sound.

Axiom & Free-Parameter Ledger

0 free parameters · 4 axioms · 3 invented entities

There are no fitted numerical parameters. The load-bearing assumptions are categorical (limits, pullbacks, monomorphism stability) and the domain-specific verifications that certain filters and sublevel sets form the required bases. The invented entities are definitions, not new physical objects, and they carry no empirical predictions of their own.

axioms (4)
  • standard math Standing categorical assumptions on C: terminal object, limits/colimits of shape S, and pullbacks when invoked.
    Used throughout Sections 2.3, 3.1-3.3, and 4.2; the paper states these assumptions as needed rather than proving them.
  • standard math Monomorphisms are closed under pullback and composition; in Set, Top, and Stoch the relevant examples satisfy this.
    Needed in Remark 4.8 and Definition 3.16 to ensure sublevel neighbourhoods are monomorphic and factorizations are unique.
  • standard math Cosieves on Out(s) form the open sets of the Alexandrov topology, and arbitrary unions and intersections of cosieves are cosieves.
    Definition 3.9 relies on this to define eventuality filters; the proof is a short categorical fact.
  • domain assumption In the examples, the tail filters on (N,≤), BN, and guarded hybrid categories are cosieve filters, and the sublevel sets form a neighbourhood basis.
    Verified only in Examples 3.18, 4.12, and 4.13; this is the point where classical topology enters the framework.
invented entities (3)
  • Abstract evolution shape S no independent evidence
    purpose: Syntax object: stages, modes, or times as objects and admissible evolutions as morphisms.
    Definition 2.1; a conceptual primitive with no external falsifiable predictions.
  • Eventuality filter E_s no independent evidence
    purpose: Formalizes the notion of 'eventually' as a filter of cosieves on outgoing morphisms.
    Definition 3.9; a definitional tool for convergence, not an empirical entity.
  • Lyapunov measurement datum (R, 0, r_lambda, <=) no independent evidence
    purpose: Abstraction of the target R_>=0 of ordinary Lyapunov functions via sublevel neighbourhoods of zero.
    Definition 4.5; a category-theoretic tool with no independent evidence beyond the examples.

reviewed 2026-08-01 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Dynamical Systems as Functorial Realisations of Abstract Evolution Shapes." pith.science (2026). https://pith.science/paper/UZPIMWCJ

@misc{pith2026260717455,
  author       = {Pith},
  title        = {Pith review of: Dynamical Systems as Functorial Realisations of Abstract Evolution Shapes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UZPIMWCJ}},
  note         = {Machine review of arXiv:2607.17455}
}
Share X Bluesky LinkedIn Reddit HN
abstract

We develop a categorical framework for closed dynamical systems in which the abstract pattern of admissible evolutions is separated from its concrete realisation. A closed dynamical system is formulated as a functor $X\colon S\to C$ from a small category $S$, viewed as an abstract evolution shape, to a coefficient category $C$. By varying $S$ and $C$, this single definition encompasses many important examples including autonomous, non-autonomous, switched, hybrid, and stochastic systems. Within this framework, we introduce invariant subsystems, equilibria, and orbits in functorial terms. We then formulate convergence by combining a cosieve-based intrinsic notion of eventuality on the evolution shape with neighbourhood filters of invariant subsystems. Finally, we establish a categorical Lyapunov principle based on categorical sublevel neighbourhoods. This yields abstract stability and convergence criteria that recover the classical Lyapunov method in standard examples.

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

19 extracted references · 10 linked inside Pith

  1. [1]

    A. D. Ames, A Categorical Theory of Hybrid Systems, Ph.D.\ thesis, University of California, Berkeley, 2006

    2006

  2. [2]

    A. D. Ames, S. Mattenet, J. Moeller, Categorical Lyapunov Theory II: Stability of Systems, arXiv:2505.22968 [math.DS]

    Pith/arXiv arXiv

  3. [3]

    A. D. Ames, J. Moeller, P. Tabuada, Categorical Lyapunov Theory I: Stability of Flows, arXiv:2502.15276 [math.DS]

    Pith/arXiv arXiv

  4. [4]

    J. C. Baez, K. Courser, Structured Cospans, Theory Appl.\ Cat.\ 35 (2020), 1771--1822; arXiv:1911.04630 [math.CT]

    Pith/arXiv arXiv 2020

  5. [5]

    J. C. Baez, K. Courser, C. Vasilakopoulou, Structured Versus Decorated Cospans, doi:10.32408/compositionality-4-3 Compositionality 4 (2022) ; arXiv:2101.09363 [math.CT]

    Pith/arXiv arXiv 2022

  6. [6]

    J. C. Baez, J. Erbele, Categories in Control, Theory Appl.\ Cat.\ 30 (2015), 836--881; arXiv:1405.6881 [math.CT]

    Pith/arXiv arXiv 2015

  7. [7]

    Behrisch, S

    M. Behrisch, S. Kerkhoff, R. Pöschel, F. M. Schneider, S. Siegmund, Dynamical Systems in Categories, doi:10.1007/s10485-015-9409-8 Appl.\ Categor.\ Struct.\ 25 (2017), 29--57

    doi:10.1007/s10485-015-9409-8 2017

  8. [8]

    M. Brin, G. Stuck, Introduction to Dynamical Systems, doi:10.1017/CBO9780511755316 Cambridge Univ.\ Press, Cambridge, 2002

    doi:10.1017/cbo9780511755316 2002

  9. [9]

    M. W. Hirsch, S. Smale, R. L. Devaney, Differential Equations, Dynamical Systems, and an Introduction to Chaos, 3rd ed., doi:10.1016/C2009-0-61160-0 Academic Press, 2013

    doi:10.1016/c2009-0-61160-0 2013

  10. [10]

    H. K. Khalil, Nonlinear Systems, Prentice Hall, Upper Saddle River, NJ, 2002

    2002

  11. [11]

    F. W. Lawvere, Taking Categories Seriously, Rev.\ Colombiana Mat.\ 20 (1986), 147--178 http://eudml.org/doc/181771

    1986

  12. [12]

    Libkind, D

    S. Libkind, D. J. Myers, Towards a Double Operadic Theory of Systems, arXiv:2505.18329 [math.CT]

    Pith/arXiv arXiv

  13. [13]

    D. J. Myers, Double Categories of Open Dynamical Systems, doi:10.4204/EPTCS.333.11 Electron.\ Proc.\ Theor.\ Comput.\ Sci.\ 333 (2021), 154--167 ; arXiv:2005.05956 [math.CT]

    Pith/arXiv arXiv 2021

  14. [14]

    D. J. Myers, Categorical Systems Theory, available on author's personal webpage https://www.davidjaz.com/Papers/DynamicalBook.pdf, 2023

    2023

  15. [15]

    J. J. M. M. Rutten, Universal Coalgebra: A Theory of Systems, doi:10.1016/S0304-3975(00)00056-6 Theoret.\ Comput.\ Sci.\ 249 (2000), 3--80

    doi:10.1016/s0304-3975(00)00056-6 2000

  16. [16]

    Schultz, D

    P. Schultz, D. I. Spivak, C. Vasilakopoulou, Dynamical Systems and Sheaves, doi:10.1007/s10485-019-09565-x Appl.\ Categor.\ Struct.\ 28 (2020), 1--57 ; arXiv:1609.08086 [math.CT]

    Pith/arXiv arXiv 2020

  17. [17]

    D. I. Spivak, J. Tan, Nesting of Dynamical Systems and Mode-Dependent Networks, doi:10.1093/comnet/cnw022 J.\ Complex Netw.\ 5 (2017), 389--408 ; arXiv:1502.07380 [math.CT]

    Pith/arXiv arXiv 2017

  18. [18]

    S. H. Strogatz, Nonlinear Dynamics and Chaos: With Applications to Physics, Biology, Chemistry, and Engineering, 2nd ed., doi:10.1201/9780429492563 CRC Press, Boca Raton, 2018

    doi:10.1201/9780429492563 2018

  19. [19]

    Vagner, D

    D. Vagner, D. I. Spivak, E. Lerman, Algebras of Open Dynamical Systems on the Operad of Wiring Diagrams, Theory Appl.\ Cat.\ 30 (2015), 1793--1822; arXiv:1408.1598 [math.CT]

    Pith/arXiv arXiv 2015

This paper was first reviewed by deepseek-v4-flash on August 1, 2026.