Pith. sign in

REVIEW 7 minor 42 references

Global linearization of asymptotically stable systems without hyperbolicity

T0 review · 0 major / 7 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read This paper proves that asymptotically stable equilibria of continuous flows admit global linearizing homeomorphisms on the whole basin of attraction, and that the conjugacy is smooth away from the equilibrium in every dimension except 5…

desk verdict Global Hartman-Grobman for asymptotically stable nonhyperbolic equilibria, complete in all dimensions for topological conjugacy and tied to the 4D smooth Poincaré conjecture in dimension 5, with sound proofs worth refereeing. read the letter →

arxiv 2502.07708 v3 pith:OBL43TTM submitted 2025-02-11 math.DS cs.SYeess.SY

Matthew D. Kvalheim , Eduardo D. Sontag This is my paper · ORCID
classification math.DScs.SYeess.SY MSC 37C1037C1537C75
keywords Hartman-GrobmantheoremnonhyperboliclinearizationasymptoticstabilityLyapunovfunctionsbasinofattractionKoopmaneigenfunctionssmoothPoincaréconjecture
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper proves that the Hartman-Grobman theorem, which classically linearizes dynamics only around hyperbolic equilibria, extends to the opposite regime: an equilibrium that is merely asymptotically stable, with no hyperbolicity assumption, can still be globally linearized. For a complete continuous vector field whose trajectories are unique, there is a homeomorphism from the entire basin of attraction to $\mathbb{R}^n$ that converts the nonlinear flow into the linear contraction $\dot y = -y$. If the flow is $C^k$ and the dimension is not $5$, the homeomorphism is actually a $C^k$-diffeomorphism away from the equilibrium. In dimension $5$, the $C^k$ statement is shown to be true if and only if the 4-dimensional smooth Poincaré conjecture is true, so the last missing regularity is tied to an open problem in topology.

What carries the argument

The load-bearing mechanism is the flow-box coordinate system built from a regular level set $L=V^{-1}(c)$ of a proper $C^\infty$ strict Lyapunov function $V$ on the basin. Every trajectory except the equilibrium crosses $L$ exactly once and transversely, so the basin minus the equilibrium is homeomorphic (and $C^k$-diffeomorphic, when the flow is $C^k$) to $\mathbb{R}\times L$ via the map $(t,x)\mapsto \Phi_t(x)$. Because $L$ is homotopy equivalent to $S^{n-1}$, and because the low- and high-dimensional classification theorems make $L$ diffeomorphic to the standard sphere for every $n\ne 5$, one can identify $L$ with $S^{n-1}$ and write the coordinate $h(x)=e^{\tau(x)}P(\rho(x))$, where $\tau$ is the arrival time and $\rho$ the crossing point. This single formula carries the argument: the linear flow on $\mathbb{R}^n$ is simply the product of exponential decay in $\mathbb{R}$ and constant motion along rays, so the conjugacy is exact and global.

What would settle it

Take a $C^\infty$ homotopy 4-sphere $L$ that is not diffeomorphic to $S^4$, build a $C^\infty$ function $V\colon S^5\to[0,1]$ having $L$ as a regular level set, and run the gradient flow $-\nabla V$. If that system admits a $C^1$ linearizing conjugacy on a neighborhood of the sink, the paper's equivalence claim is false; proving no such conjugacy for every such $L$ would confirm that the dimension-5 gap is exactly the smooth Poincaré conjecture. A failure of the underlying level-set theorem—for instance a continuous uniquely integrable flow whose regular level sets are not homotopy equivalent to spheres—would also break the construction.

Watch

Extended reading notes

Core claim

For a complete uniquely integrable continuous vector field on an $n$-dimensional smooth manifold with an asymptotically stable equilibrium $x_*$ and basin $B$, the paper proves that there is a homeomorphism $h\colon B \to \mathbb{R}^n$ satisfying $\Phi_t|_B = h^{-1}\circ e^{At}\circ h$ for every $t\in\mathbb{R}$ and every Hurwitz matrix $A$; in particular $A=-I$ turns the flow into $\dot y = -y$. If the flow is $C^k$ with $k\ge 1$ and $n\ne 5$, the same $h$ restricts to a $C^k$-diffeomorphism $B\setminus\{x_*\}\to\mathbb{R}^n\setminus\{0\}$. The paper also proves that the $C^k$ statement in dimension $n=5$ is logically equivalent to the 4-dimensional smooth Poincaré conjecture, and derives the local version as a consequence of the global one. The argument works by picking a proper strict Lyapunov function, using one regular level set as a global cross-section, and converting the time-to-reach-the-section and the position on the section into linear coordinates.

Load-bearing premise

The argument depends on the theorem that every asymptotically stable equilibrium of a complete uniquely integrable continuous flow admits a proper, strictly decreasing smooth Lyapunov function whose regular level sets are homotopy equivalent to spheres; if that theorem ever failed, the construction of the cross-section and the linearizing map collapses.

Editorial extensions

If this is right

  • Every complete asymptotically stable continuous flow is topologically conjugate, on its entire basin, to the linear system $\dot y = -y$, so the basin is homeomorphic to $\mathbb{R}^n$.
  • For $C^k$ flows in any dimension other than $5$, the linearizing coordinates are $C^k$ away from the equilibrium, preserving all finite-time derivative information up to order $k$ in the linear model.
  • Choosing $A$ diagonal produces $n$ continuous Koopman eigenfunctions whose joint map is a global homeomorphism from the basin to $\mathbb{R}^n$, giving existence of minimal-dimensional linearizing observables.
  • The equivalence with the 4-dimensional smooth Poincaré conjecture means that the remaining dimension-5 smoothness gap cannot be closed by dynamics alone; it is a question of 4-manifold topology.
  • The results provide existence targets for data-driven algorithms, such as extended dynamic mode decomposition, that seek linearizing coordinates of dimension equal to the state space.

Reading between the lines

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

  • If the construction is as robust as it appears, the same time-plus-sphere-coordinate recipe should linearize flows near any compact invariant manifold whose stable and unstable cross-sections are sphere-like, not just single equilibria.
  • The dimension-5 cliff suggests that numerical experiments claiming smooth linearization in $\mathbb{R}^5$ cannot settle the theoretical question either way: the obstruction is a global property of 4-manifolds invisible to local computation.
  • A natural testable extension is to the input-to-state setting, replacing $\dot x = f(x)$ by $\dot x = f(x,u)$ and asking whether the same Lyapunov cross-section yields finite-energy gain; the authors explicitly leave this open.
  • The one-dimensional example $h(x)=e^{-1/(2x^2)}$ for $\dot x=-x^3$ suggests that the conjugacy need not be Lipschitz at the equilibrium, so algorithms based on derivative information at the fixed point may fail even though global linearizing coordinates exist.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

0 major / 7 minor

Summary. The paper proves a global linearization theorem for asymptotically stable equilibria of nonhyperbolic systems. Theorem 1 gives a local topological conjugacy to a Hurwitz linear flow, with a C^k diffeomorphism away from the equilibrium when n≠5; Theorem 2 extends this to the whole basin of attraction for complete flows, and Proposition 1 shows that the missing n=5 C^k case is equivalent to the 4-dimensional smooth Poincaré conjecture. The proofs use a proper smooth strict Lyapunov function, the identification of the punctured basin with R×L, the classification of homotopy spheres, and Hirsch's construction.

Significance. The result is a natural and significant extension of the Hartman-Grobman theorem, resolving the global nonhyperbolic case under minimal regularity assumptions and with essentially optimal smoothness. The proof is elegant and modular, and the paper is honest about the dimensional obstruction: it proves the n=5 smooth case is equivalent to the 4-dimensional smooth Poincaré conjecture rather than claiming it. The main arguments are transparent derivations from classical external results (Wilson, Fathi–Pageault, Smale, Perelman, Freedman, Hirsch), with no fitted parameters or self-referential assumptions.

minor comments (7)
  1. [Section 3, proof of Theorem 2] The classification of the boundary L as diffeomorphic to S^{n-1} lists n=2,3,4,≥6 but omits n=1; the n=1 case is trivial because L is a two-point 0-manifold, and this should be stated for completeness.
  2. [Additional Comments, Proposition 2] This proposition is stated without a proof; the text says the proof of Theorem 2 can be repeated verbatim, but for a formal paper either a proof or an explicit label as a sketch should be supplied.
  3. [Section 3, proof of Theorem 1] The truncation function ψ is described only as equal to 1 on U0 and zero outside a compact set; to guarantee that x* is asymptotically stable for ψf with an open basin containing U0, ψ should be positive on an intermediate neighborhood, and this should be clarified.
  4. [Section 3, proof of Theorem 2] The application of Smale's theorem for n≥6 is terse; adding a sentence explaining that deleting a small ball from the contractible sublevel set V^{-1}([0,c]) yields an h-cobordism between L and S^{n-1} would make the argument self-contained.
  5. [Proposition 1] The phrase 'the C^k statement of Theorem 1 (or Theorem 2) is true for n=5' should be defined explicitly, since the main theorems exclude n=5 by hypothesis.
  6. [Abstract and title] The abstract and title contain typographical artifacts ('GLOBAL LINEARIZA TION', 'ST ABLE'); the manuscript should be copy-edited.
  7. [Section 3, proof of Theorem 2] Since the existence of a proper strict C∞ Lyapunov function for a merely continuous flow is the only non-classical external input, it would be helpful to quote the precise statement from [FP19, Sec. 6] that covers this case.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the main theorem is derived from independent external theorems (Wilson, Smale, Perelman, Freedman, Hirsch) with no step that reduces to its own conclusion.

full rationale

The paper's central derivation in Theorem 2 constructs the global linearizing homeomorphism using a proper strict C∞ Lyapunov function supplied by Wilson's theorem ([Wil69, Thm 3.2], [FP19, Sec. 6]) and then applies external topological results (Smale n≥6, Perelman n=4, Milnor/Hirsch n=2,3, Freedman n=5) to diffeomorphically trivialize the level set L. The proof does not assume the existence of the linearizing homeomorphism; it produces it explicitly as h(x) = e^{τ(x)}P(ρ(x)). The reduction from a general Hurwitz matrix A to A = −I is a standard construction that reproves the statement for the linear system, not a use of the theorem being proved. Theorem 1 is derived from Theorem 2, which is proved independently. Proposition 1 is an equivalence: the forward direction invokes the 4D smooth Poincaré conjecture to upgrade Freedman's homeomorphism to a diffeomorphism, and the converse uses Hirsch's construction to build a Lyapunov function on S^5 whose regular level set is an arbitrary homotopy sphere, then applies Theorem 1 to obtain a diffeomorphism to S^4. Neither direction assumes the theorem's conclusion for n=5. Self-citations (GSW99, KS24) appear only in historical remarks, for context about prior sketches, or for a related Proposition 2 whose proof is actually given rather than cited. No fitted parameters are renamed as predictions, and no uniqueness theorem from the authors' prior work is invoked to force the construction. The external dependencies are explicit, stated, and do not include the target results, so the derivation is self-contained modulo well-established theorems.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The central claim rests on several external theorems that are not proved in the paper: Wilson's Lyapunov function and level-set topology, the Poincaré/Smale/Freedman classification of homotopy spheres, and Hirsch's embedding of homotopy 4-spheres. There are no free parameters or invented entities.

assumptions (3)
  • domain assumption Existence of a proper C∞ strict Lyapunov function for an asymptotically stable equilibrium of a complete uniquely integrable continuous vector field, with level sets homotopy equivalent to spheres (Wilson's theorems)
    Invoked in the first paragraph of the proof of Theorem 2 to construct V and the level set L. Not proved in the paper; cited from [Wil69, Thm 3.2], [FP19, Sec. 6], and [Wil67].
  • standard math Classification of homotopy spheres: every homotopy (n-1)-sphere is diffeomorphic to S^{n-1} for n=2,3,4 (Perelman) and n≥6 (Smale); for n=5 a homeomorphism exists by Freedman
    Used in the proof of Theorem 2 to produce the diffeomorphism P: L → S^{n-1} for n≠5 and the homeomorphism for n=5.
  • standard math Hirsch's theorem: every 4-dimensional homotopy sphere occurs as a regular level set of a function on S^5 with exactly two critical points
    Used in the converse direction of Proposition 1 to realize an arbitrary homotopy 4-sphere L as a level set in S^5.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Global linearization of asymptotically stable systems without hyperbolicity." pith.science (2026). https://pith.science/paper/OBL43TTM

@misc{pith2026250207708,
  author       = {Pith},
  title        = {Pith review of: Global linearization of asymptotically stable systems without hyperbolicity},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OBL43TTM}},
  note         = {Machine review of arXiv:2502.07708}
}
abstract

We give a proof of an extension of the Hartman-Grobman theorem to nonhyperbolic but asymptotically stable equilibria of vector fields. Moreover, the linearizing topological conjugacy is (i) defined on the entire basin of attraction if the vector field is complete, and (ii) a $C^{k\geq 1}$-diffeomorphism on the complement of the equilibrium if the vector field is $C^k$ and the underlying space is not $5$-dimensional. We also show that the $C^k$ statement in the $5$-dimensional case is equivalent to the $4$-dimensional smooth Poincar\'{e} conjecture.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 30 canonical work pages

  1. [1]

    P Arathoon and M D Kvalheim, Koopman embedding and super-linearization counterexamples with isolated equilibria, arXiv preprint arXiv:2306.15126 (2023), 1--7

    work page Pith review arXiv 2023

  2. [2]

    Silverman

    V I Arnold, Ordinary differential equations, The MIT Press, 1973, Translated and Edited by Richard A. Silverman

    work page 1973

  3. [3]

    64 (2022), no

    S L Brunton, M Budi s i\' c , E Kaiser, and J N Kutz, Modern K oopman theory for dynamical systems , SIAM Rev. 64 (2022), no. 2, 229--340. 4416982

    2022

  4. [4]

    179 (2023), Paper No

    M-A Belabbas and X Chen, A sufficient condition for the super-linearization of polynomial systems, Systems Control Lett. 179 (2023), Paper No. 105588, 6. 4624015

    work page 2023

  5. [5]

    T B\'arta, R Chill, and E Fa s angov\'a, Every ordinary differential equation with a strict L yapunov function is a gradient system , Monatsh. Math. 166 (2012), no. 1, 57--72. 2901252

    work page 2012

  6. [6]

    4, 047510, 33

    M Budi s i\' c , R Mohr, and I Mezi\' c , Applied K oopmanism , Chaos 22 (2012), no. 4, 047510, 33. 3388723

    2012

  7. [7]

    J J Bramburger, Data-driven methods for dynamic systems, Society for Industrial and Applied Mathematics, Philadelphia, PA, 2024

    work page 2024

  8. [8]

    N Bhatia and G Szeg \"o , Stability theory of dynamical systems, number 161 in grundlehren der mathematischen wissenschaften, Springer-Verlag, 1970

    work page 1970

Show all 42 references
  1. [9]

    C Coleman, Local trajectory equivalence of differential systems, Proc. Amer. Math. Soc. 16 (1965), 890--892. 180752

    1965

  2. [10]

    , Addendum to: L ocal trajectory equivalence of differential systems , Proc. Amer. Math. Soc. 17 (1966), 770. 192135

    1966

  3. [11]

    9, 4202--4245

    J Eldering, M Kvalheim, and S Revzen, Global linearization and fiber bundle structure of invariant manifolds, Nonlinearity 31 (2018), no. 9, 4202--4245. 3841342

    2018

  4. [12]

    A Fathi and P Pageault, Smoothing L yapunov functions , Trans. Amer. Math. Soc. 371 (2019), no. 3, 1677--1700. 3894031

    2019

  5. [13]

    Differential Geometry 17 (1982), no

    M H Freedman, The topology of four-dimensional manifolds, J. Differential Geometry 17 (1982), no. 3, 357--453. 679066

    1982

  6. [14]

    5, 880--881

    D M Grobman, Homeomorphism of systems of differential equations, Doklady Akademii Nauk SSSR 128 (1959), no. 5, 880--881

    1959

  7. [15]

    2, 127--134

    L Gr \"u ne, E D Sontag, and F R Wirth, Asymptotic stability equals exponential stability, and ISS equals finite energy gain—if you twist your eyes , Systems & Control Letters 38 (1999), no. 2, 127--134

    1999

  8. [16]

    4, 610--620

    P Hartman, A lemma in the theory of structural stability of differential equations, Proceedings of the American Mathematical Society 11 (1960), no. 4, 610--620

    1960

  9. [17]

    Press, Princeton, NJ, 1965, pp

    M W Hirsch, On homotopy spheres of low dimension, Differential and C ombinatorial T opology ( A S ymposium in H onor of M arston M orse), Princeton Univ. Press, Princeton, NJ, 1965, pp. 199--204. 179800

    1965

  10. [18]

    33, Springer-Verlag, New York, 1994, Corrected reprint of the 1976 original

    , Differential topology, Graduate Texts in Mathematics, vol. 33, Springer-Verlag, New York, 1994, Corrected reprint of the 1976 original. 1336822

    1994

  11. [19]

    21, 18639--18663

    G Haller and B Kasz \'a s, Data-driven linearization of dynamical systems, Nonlinear Dynamics 112 (2024), no. 21, 18639--18663

    2024

  12. [20]

    W Jongeneel, Asymptotic stability equals exponential stability---while you twist your eyes, arXiv preprint arXiv:2411.03277 (2024), 1--27

    2024

  13. [21]

    M D Kvalheim and P Arathoon, Linearizability of flows by embeddings, arXiv preprint arXiv:2305.18288v6 (2024), 1--20

    2024 arXiv

  14. [22]

    J Ko and M-A Belabbas, Minimum number of observables and system invariants in super-linearization, IEEE Control Syst. Lett. 8 (2024), 2217--2222. 4816989

    2024

  15. [23]

    D 425 (2021), Paper No

    M D Kvalheim and S Revzen, Existence and uniqueness of global K oopman eigenfunctions for stable fixed points and periodic orbits , Phys. D 425 (2021), Paper No. 132959, 20. 4275046

    2021

  16. [24]

    M D Kvalheim and E D Sontag, Why should autoencoders work?, Transactions on Machine Learning Research (2024)

    2024

  17. [25]

    D 242 (2013), 42--53

    Y Lan and I Mezi\' c , Linearization in the large of nonlinear systems and K oopman operator spectrum , Phys. D 242 (2013), 42--53. 3001394

    2013

  18. [26]

    Z Liu, N Ozay, and E D Sontag, On the non-existence of immersions for systems with multiple omega-limit sets, IFAC World Congress, Yokohoma, Japan 56 (2023), 60--64

    2023

  19. [27]

    Preprint can be found in https://arxiv.org/abs/2312.17045, 2023/2024

    , Properties of immersions for systems with multiple limit sets with implications to learning K oopman embeddings , Automatica (2025), To appear. Preprint can be found in https://arxiv.org/abs/2312.17045, 2023/2024

    2025 arXiv

  20. [28]

    Nonlinear Sci

    I Mezi\'c, Spectrum of the K oopman operator, spectral expansions in functional spaces, and state-space geometry , J. Nonlinear Sci. 30 (2020), no. 5, 2091--2145. 4163461

    2020

  21. [29]

    J Milnor, Morse theory, Annals of Mathematics Studies, vol. No. 51, Princeton University Press, Princeton, NJ, 1963, Based on lecture notes by M. Spivak and R. Wells. 163331

    1963

  22. [30]

    Weaver, Revised reprint of the 1965 original

    J W Milnor, Topology from the differentiable viewpoint, Princeton Landmarks in Mathematics, Princeton University Press, Princeton, NJ, 1997, Based on notes by David W. Weaver, Revised reprint of the 1965 original. 1487640

    1997

  23. [31]

    18, American Mathematical Society, Providence, RI, 1996, Reprint of the 1932 original

    M Morse, The calculus of variations in the large, American Mathematical Society Colloquium Publications, vol. 18, American Mathematical Society, Providence, RI, 1996, Reprint of the 1932 original. 1451874

    1996

  24. [32]

    3, American Mathematical Society, Providence, RI; Clay Mathematics Institute, Cambridge, MA, 2007

    J Morgan and G Tian, R icci flow and the P oincar\'e conjecture , Clay Mathematics Monographs, vol. 3, American Mathematical Society, Providence, RI; Clay Mathematics Institute, Cambridge, MA, 2007. 2334563

    2007

  25. [33]

    G Perelman, The entropy formula for the R icci flow and its geometric applications , arXiv preprint math/0211159 (2002), 1--39

    2002 arXiv

  26. [34]

    , Finite extinction time for the solutions to the R icci flow on certain three-manifolds , arXiv preprint math/0307245 (2003), 1--7

    2003 arXiv

  27. [35]

    , Ricci flow with surgery on three-manifolds, arXiv preprint math/0303109 (2003), 1--22

    2003 arXiv

  28. [36]

    S Smale, On the structure of manifolds, Amer. J. Math. 84 (1962), 387--399. 153022

    1962

  29. [37]

    D eterministic F inite- D imensional S ystems , second ed., Texts in Applied Mathematics, vol

    E D Sontag, Mathematical C ontrol T heory. D eterministic F inite- D imensional S ystems , second ed., Texts in Applied Mathematics, vol. 6, Springer-Verlag, New York, 1998

    1998

  30. [38]

    Willems, S

    E D Sontag, Contractive systems with inputs, Perspectives in Mathematical System Theory, Control, and Signal Processing (J. Willems, S. Hara, Y. Ohta, and H. Fujioka, eds.), Springer-verlag, 2010, pp. 217--228

    2010

  31. [39]

    Differential Equations 3 (1967), 323--329

    F W Wilson, Jr, The structure of the level surfaces of a L yapunov function , J. Differential Equations 3 (1967), 323--329. 0231409

    1967

  32. [40]

    , Smoothing derivatives of functions and applications, Trans. Amer. Math. Soc. 139 (1969), 413--428. 0251747

    1969

  33. [41]

    C onf., N orth D akota S tate U niv., F argo, N

    , A reformulation of C oleman's conjecture concerning the local conjugacy of topologically hyperbolic singular points , The structure of attractors in dynamical systems ( P roc. C onf., N orth D akota S tate U niv., F argo, N . D ., 1977), Lecture Notes in Math., vol. 668, Spr...

    1977

  34. [42]

    M O Williams, I G Kevrekidis, and C W Rowley, A data--driven approximation of the K oopman operator: Extending dynamic mode decomposition , Journal of Nonlinear Science 25 (2015), 1307--1346

    2015

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.