Pith. sign in

REVIEW 2 major objections 4 minor 20 references

Residual pathologies

T0 review · 2 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read The paper proves that, in three classical settings, the pathological behavior is residual—typical in the Baire-category sense—rather than an isolated exception.

desk verdict Solid Baire-genericity upgrades of three known counterexamples, with one genuinely omitted approximation argument in the wave section that needs repair before publication. read the letter →

arxiv 1908.09496 v2 pith:XY5R4GM7 submitted 2019-08-26 math.FA math.AP

Marina Ghisi , Massimo Gobbino This is my paper · ORCID
classification math.FAmath.AP MSC 26A2435L1535Q3576F25
keywords BairecategoryresidualsetapproximatedifferentiabilityHöldercontinuityGevreyspacesderivativelosswaveequationtransport
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

Three famous counterexamples in analysis are usually read as rare "perfect storms" built by delicate iteration: a $C^1$ function whose derivative is Hölder continuous and approximately differentiable almost everywhere but never agrees with any $C^{1,1}$ function on a set of positive measure; a wave equation with Hölder-continuous propagation speed whose smooth initial data instantly lose all derivatives; and a transport equation with non-Lipschitz divergence-free velocity whose smooth initial data lose all Sobolev regularity for every positive time. This paper proves that in each case the pathological behavior is generic: the "bad" objects form a residual set in a natural complete metric space, so a random admissible object is overwhelmingly likely to be pathological. The argument is a Baire-category template: encode the non-pathological objects as a countable union of closed sets with empty interior, and verify emptiness of interior by rescaling a smooth basic ingredient inside any ball. The paper also identifies open problems at the critical thresholds where the generic picture is not yet known.

What carries the argument

The load-bearing mechanism is the Baire category theorem applied to spaces defined by quantitative stability conditions. For each problem the "good" objects are written as a countable union of sets $C_k$ that encode a quantitative bound: a bounded interval where $f$ agrees with a bounded $C^{1,1}$ function on a fixed fraction of small intervals; a bounded time where the wave solution lies in $G^{-B,k}(A)$ with norm at most $k$; a bounded time where the transport solution has $\|\rho(t)\|_{H^{1/k}}\le k$. Each $C_k$ is closed, and its interior is shown empty by a three-step perturbation: regularize the center of a ball, add a small rescaling of a "basic ingredient" that violates the quantitative bound while staying inside the ball, and let the scale go to infinity. The basic ingredients are the multibump functions $\phi_{n,k}$ with the gap estimate (2.21); the family $w(\varepsilon,t)=\sin t\,\exp(\varepsilon(2t-\sin 2t))$, which solves $w''+\gamma(\varepsilon,t)w=0$ with exponential growth and, rescaled, gives growth $\exp(\mathrm{const}\cdot \lambda^{1-\alpha}t)$ in the wave problem; and a quoted smooth velocity/datum pair whose transported density has homogeneous Sobolev norms growing exponentially in time.

What would settle it

Find one open ball in any of the three spaces whose elements all satisfy the quantitative non-pathological condition—for example pairs $(u,\theta)$ with $\|\rho(t)\|_{H^{1/k}}\le k$ for some $t\in[1/k,k]$. The paper proves each such set $C_k$ is closed and has empty interior, so a single such ball with nonempty interior would refute the residual genericity theorem for that problem.

Watch

Extended reading notes

Core claim

The central discovery is structural: none of these three pathologies is a property of a specially constructed object; each is the default behavior of a generic element of the appropriate space. In the notation of the paper, Theorem 2.1 shows every $f$ in a complete metric space $X$ of $C^1$ functions with nonnegative, $\alpha$-Hölder, locally Lipschitz-on-$K_n$ derivatives has $f'$ approximately differentiable almost everywhere, and a residual subset has the stronger property that for every $C^{1,1}$ function $g$ the coincidence set $\{x:f(x)=g(x)\}$ has measure zero. Theorem 3.2 shows that for every pair $(\alpha,\beta,B)$ in the regime $\beta>1/(1-\alpha)$, $B>1/(1-\alpha)$, a residual set of pairs $(c,\psi)$—a Hölder-continuous propagation speed and an initial velocity in a Gevrey class $G^{\beta,\infty}(A)$—produces a wave-equation solution that for every $t>0$ and every $R>0$ lies outside the Gevrey ultradistribution space $G^{-B,R}(A)\times G^{-B,R}(A)$ (a scale of very rough distributions associated with the operator $A$). Theorem 4.1 shows that a residual set of pairs $(u,\theta)$, where $u$ is a compactly supported divergence-free velocity with $W^{1,p}$ regularity for every $p<\infty$ and $\theta$ is $C^\infty$ and compactly supported, yields a transport solution $\rho(t)$ that belongs to no Sobolev space $H^s(\mathbb{R}^d)$ for any $t>0$, $s>0$.

Load-bearing premise

For the transport result, the paper assumes without proof the existence of a smooth divergence-free velocity field and a smooth compactly supported initial datum whose solution's homogeneous Sobolev norms grow exponentially in time; if no such pair existed, Theorem 4.1 would have no basic ingredient and the residual derivative-loss argument would collapse.

Editorial extensions

If this is right

  • In each of the three problems the non-pathological objects form a meager set, so arbitrary small perturbations of a good object are enough to reach a pathological one.
  • For the wave equation, severe derivative loss occurs for a residual set of propagation speeds and initial velocities, even when the speed is confined to a small Hölder neighborhood and the initial datum is fixed to zero.
  • For the transport equation, residual many admissible pairs produce a solution that for every positive time lies outside every Sobolev space $H^s$.
  • For approximate differentiation, residual many $f$ satisfy the stronger condition (A4-s), so no $C^{1,1}$ function can coincide with $f$ on a set of positive measure.

Reading between the lines

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

  • The same Baire-category template suggests that other counterexamples built by rescaling a smooth basic ingredient—for instance the degenerate wave equations treated in the same circle of ideas—should have residual versions, with the quantitative sets defined by the natural energy or Gevrey norms.
  • A testable extension for the transport problem: if one can strengthen the quoted exponential-growth theorem to growth of $\|\rho_*(t)\|_{H^s}$ inside every open subset of the support, then the residual derivative loss should localize in any open subset, settling the paper's Open Problem 3.
  • The residual viewpoint has a practical consequence for numerical experiments: a randomly generated velocity field in $V$ and a random smooth compactly supported datum should show Sobolev norms of the transported density growing without bound; failure of such growth in an open range of random choices would indicate an extra hidden constraint not present in the space $X$.
  • For the wave problem, the same approach likely applies to the critical Gevrey index $s=(1-\alpha)^{-1}$ with finite radius; whether derivative loss beyond a finite time is generic there is the paper's Open Problem 2, and the residual machinery may be the natural tool.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 4 minor

Summary. The paper establishes three Baire-category genericity theorems for classical pathologies. In Section 2, Theorem 2.1 shows that in a complete metric space of C^1 functions whose derivatives are nonnegative, globally alpha-Hölder, and Lipschitz on a nested family of closed sets K_n, every function has an approximately differentiable derivative almost everywhere, and a residual subset has the stronger coincidence property (A4-s): it agrees with any C^{1,1} function only on a Lebesgue-null set. In Section 3, Theorem 3.2 shows that for the abstract wave equation u''+c(t)Au=0, with c in a Hölder space F and initial velocity in a Gevrey-type space G^{β,∞}, a residual set of pairs (c,ψ) exhibits the maximal derivative loss (3.7) at every positive time. In Section 4, Theorem 4.1 shows that for the transport equation with divergence-free velocity fields in a complete space V and C^∞ compactly supported initial data in F, a residual set of pairs (u,θ) loses all Sobolev regularity at every positive time. The method is uniform: encode a quantitative version of the non-pathological behavior as a countable union of closed sets C_k, prove each C_k has empty interior, regularize the center, and insert a rescaled basic ingredient to contradict the quantitative bound. Each section also states an open problem.

Significance. If the proofs are completed as indicated below, the paper makes a valuable methodological contribution by showing that three well-known pathological counterexamples are not isolated but are generic in suitable complete metric spaces. The Baire-category framework is presented clearly, and the qualitative-versus-quantitative distinction is effective. Section 2 is essentially self-contained and contains a careful approximation lemma (Lemma 2.6) that is a model for what is missing in Section 3. Section 4 makes honest use of the external Theorem D from Alberti-Crippa-Mazzucato and the DiPerna-Lions stability lemma, explicitly identifying the black-box input; the scaling computation in (4.13)-(4.16) is coherent. The paper is also honest about its limits, including three open problems. The main obstacles are local to Section 3: one omitted density proof for the regularization of the center, and an apparent scaling error in the verification that the modified propagation speed stays in F. Both appear fixable without changing the structure of the paper.

major comments (2)
  1. [Section 3.4] The proof that 'The set C_k has empty interior' begins with the assertion, made via 'Up to a small modification', that the center (c0,ψ0) can be assumed Lipschitz continuous, constant on [0,δ] for some δ in (0,1/k0), and satisfying the strict inequalities (3.20)-(3.21). No proof of this density statement is given. The assertion is load-bearing: the construction of c_n via (3.22) needs c0(t)=m^2 on [0,δ_n] so that c_n is continuous at δ_n and the explicit WKB-type solution is valid on [0,δ_n]; Lemma 3.6 requires c0 to be Lipschitz on [δ_n,+∞); and the contradiction requires the modified center to remain in F with the same Hölder constant H, so that a whole ball is still contained in C_{k0}. Since the analogous approximation step is proved carefully for the space X in Section 2 (Lemma 2.6) but omitted here, this is not merely a stylistic detail. Please supply a proof or a precise reference establishing that such regularized centers are dense in F.
  2. [Section 3.4, equations (3.11) and (3.22)] The verification that c_n belongs to F does not follow from the stated definition of ε_n. According to (3.11), the α-Hölder oscillation of c(t)=m^2 γ(ε,mλt) around the constant value m^2 is controlled by ε m^2 (mλ)^α Hγ |t-s|^α. With ε_n defined in (3.22) as ε_1 H (m^α+2Hγ)^{-1} λ_n^{-α}, the resulting Hölder constant for c_n-c0 is ε_1 H · Hγ m^{2+α}/(m^α+2Hγ), which need not be bounded by ε_1 H; when μ_2 is large, so that m is large, this quantity can exceed ε_1 H. Therefore the conclusion that c_n-c0 has Hölder constant ε_1 H, and hence that c_n∈F, is not justified. Please correct the scaling in (3.22) or explain the intended additional normalization of Hγ; this point is load-bearing for the claim that (c_n,ψ_n)∈B_X((c0,ψ0),ε0).
minor comments (4)
  1. [Section 4.4, condition (4.14)] The parameters α and β in (4.14) are fresh small parameters, but they reuse the notation α,β from the earlier statements of Theorem B and Theorem 3.2, where they have a completely different meaning. Please rename them, for example a and b, to avoid confusion.
  2. [Section 4.4, regularization of the center] The three-step approximation of (u0,θ0) is described in words and is plausible, but the effect of the initial rescaling on the norm bound (4.7) is not written out. A short verification that convolution does not enlarge the support beyond B(0,1) and preserves the p^4 bound would improve the exposition.
  3. [Throughout] There are several minor typographical errors, including 'sp ecial' in the abstract and 'cathegory' in the acknowledgments; these should be corrected during revision.
  4. [Section 3.4, closure of C_k] The closure argument for C_k uses lower semicontinuity of the G^{-B,k}(A)-norm under componentwise convergence; this is correct, but a one-line justification via Fatou's lemma would make the argument easier for the reader to check.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the three residual-genericity theorems are derived from explicit external basic ingredients (Kohn, Colombini-De Giorgi-Spagnolo, Alberti-Crippa-Mazzucato), not from their own conclusions.

full rationale

The three Baire-category results are not equivalent to their inputs by construction. Section 2 uses the Kohn multi-bump functions (2.5)-(2.6) only as perturbations to prove C_k has empty interior; the residual A4-s conclusion is not built into the definition of X. Section 3 uses the explicit WKB-type solution w(ε,t) from [9] to make small perturbations c_n, ψ_n of an arbitrary center and derives the lower bound by the energy Lemma 3.6; the derivative-loss set is not defined to contain these perturbations. Section 4 imports Theorem D from [1,2] as an external black box (a deep smooth example with exponential homogeneous-Sobolev growth), then rescales it to displace any ball around an arbitrary (u0,θ0); Theorem 4.1 is not the same statement as Theorem D, and the genericity claim is added by the Baire argument. Self-citations [14,15,16] are used only for standard definitions and as references for extensions, not as the load-bearing premise. The only flagged issue is a proof omission, not circularity: in Section 3.4, 'Regularization of the center' asserts 'Up to a small modification of c0, and a small reduction of the radius ε0, we can assume that c0 has the following further properties' (Lipschitz, constant near 0, non-saturating); this density/approximation statement is load-bearing for the definition of c_n and for Lemma 3.6, and it is not proved in the paper. It is a completeness gap that could be repaired by a standard mollification plus time-shift, and it does not make any theorem reduce to its own inputs. Thus no claim in the paper is circular in the sense of the requested analysis.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The central claims rest on standard Baire category plus explicit constructions. The wave and differentiation sections provide self-contained proofs of the needed basic ingredients; the transport section relies on the deep prior theorem of Alberti-Crippa-Mazzucato as a black box. No free parameters are fitted to data; all constants are chosen to satisfy explicit inequalities. No new entities are introduced.

assumptions (5)
  • standard math Baire category theorem
    Used to conclude that a countable intersection of dense open sets is residual; all three proofs show the complement of the pathological set is a countable union of closed sets with empty interior. Invoked throughout Section 1.
  • standard math Completeness of the chosen function spaces X under their metrics
    Baire theorem requires complete metric spaces; the spaces X in Sections 2-4 are shown to be complete, but this is a structural assumption underpinning genericity.
  • domain assumption Stability of transport solutions (DiPerna-Lions), Lemma 4.4
    Used to show that the non-pathological set C_k is closed in the transport section; it is quoted from [12, Theorem II.4] and not reproved.
  • domain assumption Theorem D of Alberti-Crippa-Mazzucato (basic ingredient for transport)
    Asserts existence of a smooth divergence-free velocity field and smooth initial data whose solution has exponential growth of homogeneous Sobolev norms; used as the rescaled perturbation in Theorem 4.1. Stated in Section 4.1.
  • standard math Explicit ODE solution w(ε,t) = sin t exp(ε(2t - sin 2t)) solving w'' + γw = 0
    Direct computation provides the basic ingredient for the wave equation; used in Sections 3.3 and 3.4.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Residual pathologies." pith.science (2026). https://pith.science/paper/XY5R4GM7

@misc{pith2026190809496,
  author       = {Pith},
  title        = {Pith review of: Residual pathologies},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XY5R4GM7}},
  note         = {Machine review of arXiv:1908.09496}
}
read the original abstract

Several counterexamples in analysis show the existence of some special object with some sort of pathological behavior. We present three different examples where the pathological behavior is not an isolated exception, but it is the "typical" behavior of the "generic" object in a suitable class, where here generic means residual in the sense of Baire category. The first example is the revisitation of a classical result concerning approximate differentiation. The second example is the derivative loss for solutions to linear wave equations with time-dependent Holder continuous propagation speed. The third result is the derivative loss for solutions to transport equations with non-Lipschitz velocity field.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

20 extracted references · 20 canonical work pages

  1. [1]

    Alberti, G

    G. Alberti, G. Crippa, A. L. Mazzucato . Exponential self-similar mixing by incompressible flows. J. Amer. Math. Soc. 32 (2019), no. 2, 445–490

    work page 2019

  2. [2]

    Alberti, G

    G. Alberti, G. Crippa, A. L. Mazzucato . Loss of regularity for the continuity equation with non-Lipschitz velocity field. Ann. PDE 5 (2019), no. 1, Art. 9, 19

    work page 2019

  3. [3]

    Ambrosio

    L. Ambrosio . Transport equation and Cauchy problem for BV vector fields. Invent. Math. 158 (2004), no. 2, 227–260

    work page 2004

  4. [4]

    Ambrosio , G

    L. Ambrosio , G. Crippa. Continuity equations and ODE flows with non-smooth velocity. Proc. Roy. Soc. Edinburgh Sect. A 144 (2014), no. 6, 1191–1244

    work page 2014

  5. [5]

    Bahouri , J.-Y

    H. Bahouri , J.-Y. Chemin , R. Danchin . Fourier analysis and nonlinear par- tial differential equations , Grundlehren der Mathematischen Wissenschaften [Fun- damental Principles of Mathematical Sciences] , volume 343. Springer, Heidelberg, 2011

    work page 2011

  6. [6]

    J. M. Borwein , W. B. Moors , X. W ang. Generalized subdifferentials: a Baire categorical approach. Trans. Amer. Math. Soc. 353 (2001), no. 10, 3875–3893

    work page 2001

  7. [7]

    Sharp regularity estimates for solutions of the continuity equation drifted by Sobolev vector fields

    E. Bru` e, Q.-H. Nguyen. Sharp regularity estimates for solutions of the continuity equation drifted by sobolev vector fields. arXiv:1806.03466

    work page Pith review arXiv

  8. [8]

    A. Cellina . On the differential inclusion x′ ∈ [−1, +1]. Atti Accad. Naz. Lincei Rend. Cl. Sci. Fis. Mat. Nat. (8) 69 (1980), no. 1-2, 1–6 (1981)

    work page 1980

Show all 20 references
  1. [9]

    Colombini , E

    F. Colombini , E. De Giorgi , S. Spagnolo . Sur les ´ equations hyperboliques avec des coefficients qui ne d´ ependent que du temps. Ann. Scuola Norm. Sup. Pisa Cl. Sci. (4) 6 (1979), no. 3, 511–559

    1979

  2. [10]

    Colombini, E

    F. Colombini, E. Jannelli , S. Spagnolo. Well-posedness in the Gevrey classes of the Cauchy problem for a nonstrictly hyperbolic equation with coe fficients de- pending on time. Ann. Scuola Norm. Sup. Pisa Cl. Sci. (4) 10 (1983), no. 2, 291–312

    1983

  3. [11]

    Dacorogna, P

    B. Dacorogna, P. Marcellini. General existence theorems for Hamilton-Jacobi equations in the scalar and vectorial cases. Acta Math. 178 (1997), no. 1, 1–37

    1997

  4. [12]

    R. J. DiPerna , P.-L. Lions . Ordinary differential equations, transport theory and Sobolev spaces. Invent. Math. 98 (1989), no. 3, 511–547

    1989

  5. [13]

    H. Federer . Geometric measure theory . Die Grundlehren der mathematischen Wissenschaften, Band 153. Springer-Verlag New York Inc., New Yo rk, 1969

    1969

  6. [14]

    Ghisi, M

    M. Ghisi, M. Gobbino. Derivative loss for Kirchhoff equations with non-Lipschitz nonlinear term. Ann. Sc. Norm. Super. Pisa Cl. Sci. (5) 8 (2009), no. 4, 613–646. 36

    2009

  7. [15]

    Ghisi , M

    M. Ghisi , M. Gobbino . Linear wave equations with time-dependent propagation speed and strong damping. J. Differential Equations 260 (2016), no. 2, 1585–1621

    2016

  8. [16]

    Ghisi , M

    M. Ghisi , M. Gobbino . Time-dependent propagation speed vs strong damping for degenerate linear hyperbolic equations. J. Differential Equations 266 (2019), no. 1, 114–146

    2019

  9. [17]

    Kirchheim , E

    B. Kirchheim , E. Spadaro , L. Sz ´ ekelyhidi, Jr. Equidimensional isometric maps. Comment. Math. Helv. 90 (2015), no. 4, 761–798

    2015

  10. [18]

    R. V. Kohn . An example concerning approximate differentiation. Indiana Univ. Math. J. 26 (1977), no. 2, 393–397

    1977

  11. [19]

    T. W. K ¨orner. Besicovitch via Baire. Studia Math. 158 (2003), no. 1, 65–78

    2003

  12. [20]

    U. Menne . The concept of varifold. Notices Amer. Math. Soc. 64 (2017), no. 10, 1148–1152. 37

    2017

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.