Pith. sign in

REVIEW 3 minor 14 references

Lions' Maximal Regularity Problem for Divergence-Form Differential Operators: Failure at the $\frac{1}{2}$-H\"older Endpoint

T0 review · 0 major / 3 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Even a 1/2-Hölder coefficient arbitrarily close to 1 can make the heat equation's unique solution lose square-integrable time regularity.

desk verdict Genuine endpoint counterexample for scalar divergence-form operators with C^{0,1/2} coefficients; explicit, self-contained, and convincing. read the letter →

arxiv 2608.11194 v1 pith:V7UNHFWU submitted 2026-08-11 math.AP math.FA

Lukas Niebel This is my paper · ORCID
classification math.APmath.FA MSC 35K9035B6535R05
keywords non-autonomousmaximalregularityLions'problemdivergence-formoperators1/2-Hölderendpointcounterexamplevariationalsolutionlacunaryfrequenciesboundeddomains
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

This paper settles Lions' maximal-regularity problem at the critical half-Hölder endpoint. It constructs, on the interval $(0,\pi)$, a uniformly elliptic real scalar diffusion coefficient that is $1/2$-Hölder continuous in time with values in $L^\infty$ and can be chosen arbitrarily close to the constant coefficient $1$. For zero initial data and a forcing term that is continuous in time and square-integrable in space, the unique variational solution has a time derivative that is not square-integrable in space-time. This proves that $C^{0,1/2}$ time regularity alone does not imply maximal $L^2$-regularity for divergence-form operators, confirming a conjecture discussed in the paper. Zero extension, tensorisation, and parabolic rescaling turn the interval counterexample into real symmetric isotropic counterexamples on $\mathbb{R}^d$ and on every bounded domain, for all $d\ge1$.

What carries the argument

The engine is the coefficient–profile identity $L_{K^{-1}\cos(Kx)}\phi = \sin(Kx)\sin(2x) - 2K^{-1}\cos(Kx)\cos(2x)$, with $\phi(x)=\sin^2 x$. It says that a coefficient oscillation of amplitude $K^{-1}$ and spatial frequency $K$, differentiated once, produces a leading mode whose $L^2$ norm is independent of $K$; the leftover term carries an extra factor $K^{-1}$. The coefficient modulates this at temporal frequency $K^2$, which matches parabolic scaling and gives exactly the factor $K^{-1}\min\{2,K^2|t-s|\}\le\sqrt2|t-s|^{1/2}$, so the coefficient is $1/2$-Hölder and no better. Lacunary spacing $K_{j,m+1}=16K_{j,m}$ makes the modes orthogonal across blocks, so the large contributions to $\partial_t u$ add up instead of cancelling; scalar amplitudes $y_{j,m}$ satisfying $\partial_t y_{j,m}+K_{j,m}^2 y_{j,m}$ are tuned to cancel the leading mode in the forcing, keeping $f$ in $C([0,1];H)$ while the time derivative accumulates a divergent harmonic series.

What would settle it

For the explicitly constructed coefficient and forcing, compute or simulate with the first N blocks the integral $\int_0^1 \|\partial_t u(t)\|_{L^2(0,\pi)}^2\,dt$. The paper shows this is bounded below by a constant times $\sum_{j=1}^N M_j\ell_j$, which grows like the harmonic series; if a careful high-resolution computation found the partial sums bounded, the claimed divergence—and hence the theorem—would be false.

Watch

Extended reading notes

Core claim

The central claim is Theorem 1.1: for every $\delta>0$ there exist a uniformly elliptic real coefficient $a\in C^{0,1/2}([0,1];L^\infty(0,\pi))$ with $\|a-1\|_{C^{0,1/2}}<\delta$ and a forcing $f\in C([0,1];L^2)\cap L^2$ such that the unique Lions variational solution $u$ with $u(0)=0$ satisfies $u(t)\in D(A_H(t))$ for every $t$, yet $\partial_t u\notin L^2(0,1;H)$ and $A_H(\cdot)u(\cdot)\notin L^2(0,1;H)$, where $H=L^2(0,\pi)$. The coefficient is formed by superposing oscillations at spatial frequency $K$ and temporal frequency $K^2$ on disjoint time blocks of lengths $\ell_j\simeq 1/(16j(j+1))$, with $j$ modes on the $j$-th block. On block $j$ the squared $L^2$ norm of $\partial_t u$ is comparable to $\varepsilon^2/(16(j+1))$, so the total diverges like a harmonic series even though the forcing is continuous with values in $H$.

Load-bearing premise

The counterexample depends on the lacunary separation of the oscillatory modes: consecutive spatial frequencies differ by a factor of 16, which makes the mode families orthogonal and prevents the bad time-derivative contributions from cancelling across blocks; if that separation were removed, cross-mode interaction could restore square integrability.

Editorial extensions

If this is right

  • At the critical Hölder exponent $1/2$, no general maximal $L^2$-regularity theorem can hold for scalar divergence-form operators on intervals, because the coefficient here is real and uniformly elliptic and can be taken arbitrarily close to $1$.
  • The counterexample transfers to the full space $\mathbb{R}^d$ and to every bounded domain $\Omega\subset\mathbb{R}^d$ for all $d\ge1$ with a real symmetric isotropic coefficient matrix, so the failure is neither a boundary artefact nor a low-dimensional phenomenon.
  • The construction confirms that the sufficient endpoint conditions in the paper's cited positive results—bounded variation, Dini-type moduli, piecewise $H^{1/2}$, and the scale-invariant square condition—are all genuinely needed; the constructed coefficient manages to violate each of them.
  • Maximal regularity fails despite the solution being well-behaved pointwise: $u(t)$ lies in the operator domain $D(A_H(t))$ for every $t$ and $A_H(\cdot)u(\cdot)$ belongs to $L^1$, but the square-integrable regularity of the time derivative is exactly what is lost.
  • The same construction gives a coefficient that is $1/2$-Hölder with arbitrarily small norm but not $C^{0,\alpha}$ for any $\alpha>1/2$, so the endpoint scaling is sharp.

Reading between the lines

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

  • A testable extension is whether a similar lacunary construction can be built for quasilinear or higher-order parabolic problems; the load-bearing identity is specific to second-order divergence form, so the failure mechanism would need to be re-derived there.
  • The construction suggests that any sufficient condition for endpoint maximal regularity must be non-local in time and must couple spatial structure, because pointwise $1/2$-Hölder continuity with values in $L^\infty$ is not enough even at arbitrarily small amplitude.
  • A quantitative version may be possible: truncating the construction at $N$ blocks should make the squared $L^2$ norm of $\partial_t u$ grow like $\log N$, giving a concrete rate at which regularity degrades as the counterexample is approximated.
  • Because the flux vanishes at the endpoints, the zero-extension trick is robust; a similar profile with vanishing first derivatives could be sought for other boundary conditions, though such an extension is not part of the paper.
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 / 3 minor

Summary. The paper gives a counterexample to Lions' maximal L^2-regularity problem for divergence-form operators on a bounded interval: for every δ>0 it constructs a real uniformly elliptic coefficient a∈C^{0,1/2}([0,1];L∞(0,π)) with ||a−1||_{C^{0,1/2}}<δ, and an H-valued continuous forcing f, such that the unique Lions variational solution has ∂_t u∉L^2(0,1;H). The construction superposes lacunary oscillatory modes p_K=sin(Kx)sin(2x) on shrinking time blocks, with temporal frequency K^2 and amplitude K^{-1}; a scalar corrector y_{j,m} cancels the leading p_K contributions so that the forcing is continuous in H, while the norm of ∂_t u on the blocks contains a divergent harmonic series. The authors then extend the construction by zero to the real line, tensorise it to R^d, and localise it by parabolic rescaling to arbitrary bounded domains, and they compare the coefficient with the known endpoint sufficient conditions in the literature.

Significance. If the construction is correct, it settles the endpoint question for Lions' maximal regularity in the divergence-form setting: C^{0,1/2} time regularity alone is insufficient, even for real scalar coefficients that are arbitrarily small perturbations of the identity. This confirms Auscher–Egert's conjecture and sharpens earlier abstract counterexamples by Fackler and Dier. The proof is explicit and self-contained; the key identities (Lemma 2.1), the cancellation (5.2), and the divergent block sum (Lemma 4.1) are verified with estimates that carry no fitted parameters. The passage to R^d and to bounded domains is a genuine extension of the interval construction, not a formal modification, and the comparison with sufficient hypotheses in Section 7 clarifies the precise position of the example.

minor comments (3)
  1. [§4, Lemma 4.1] The proof of the C^1 regularity of v cites [2, Chap. V, §2, Thm. 2.8, p. 373] for the termwise differentiation of a uniformly convergent series of V'-valued functions; this reference appears to be an ordinary analysis textbook rather than a Bochner-space reference, and the pagination may not correspond to the cited edition. The argument itself is standard, but the citation should be checked and, if necessary, replaced by a standard reference on vector-valued differentiation.
  2. [§3, Proposition 3.1] In the proof of the 1/2-Hölder estimate, the step 'Ch+Ch^{1/2} ≤ Ch^{1/2}' uses h≤ℓ_j≤1 implicitly; making this explicit would improve readability, especially in the cross-block case where the same inequality is used with h>ℓ_j.
  3. [§7, Corollary 7.1] The exclusions in Corollary 7.1 are all proved by contradiction from known positive results, which is logically sufficient. A short direct indication of why the constructed coefficient fails the Auscher–Egert square condition (7.1), for instance by the same Fourier-coefficient argument as in Proposition 3.2, would make the comparison more transparent for readers who do not immediately see the divergence of the integral.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the counterexample is constructed explicitly and its failure of maximal regularity is proved from direct estimates.

full rationale

The paper is a self-contained counterexample construction rather than a prediction from fitted inputs. The coefficient a is built from an explicit lacunary superposition (Section 3), the candidate solution u is assembled from explicit amplitudes and modes (Section 4), and the forcing f is then defined by f = dt u + A(t)u (equation (5.3)). Defining f as the residual is legitimate in an existence proof and does not smuggle in the target conclusion. The failure of maximal L2-regularity is derived independently in Lemma 4.1: on the active blocks the time derivative has the explicit form (4.7), and the orthogonality of the lacunary modes (Lemma 2.1), together with the harmonic-series lower bound sum M_j ell_j = 1/(16(j+1)), gives a divergent squared L2 norm. No parameter is fitted to a dataset and then called a prediction. The proof does not assume that maximal regularity fails; the contradiction comes from the explicit lower bound. The cited endpoint results are used only for context and for the post-hoc comparisons in Section 7, and none of those citations is load-bearing for the construction. There are no self-citations by the author, no imported uniqueness theorem, and no renaming of a known empirical pattern. The derivation chain is therefore non-circular, and the appropriate score is 0.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests entirely on an explicit construction. The only hand-chosen numerical parameters are the small amplitude epsilon, the lacunarity ratio 16, and the block-length schedule ell_j; none are fitted to data or to the conclusion. The analytic inputs are standard theorems. No new entities (particles, forces, dimensions) are postulated.

free parameters (3)
  • epsilon (perturbation amplitude) = 0 < epsilon < delta/C
    Controls the size of the coefficient perturbation a = 1 + epsilon*beta; chosen small enough to ensure uniform ellipticity and ||a-1||_{C^{0,1/2}} < delta.
  • lacunarity base 16 = 16
    Sets K_{j,m+1} = 16 K_{j,m}, making frequency tails geometric in Proposition 3.1 and ensuring orthogonality (L >= 16K) in Lemma 2.1.
  • block length schedule ell_j = 1/(16 j(j+1))
    Chosen so M_j ell_j = 1/(16(j+1)) diverges harmonically while sum ell_j and sum M_j S_j converge, producing the failure of L2-integrability.
assumptions (4)
  • standard math Lions' variational theorem (existence and uniqueness of variational solutions)
    Invoked in Section 1 and in the proofs of Theorem 1.1, Proposition 6.2, and Corollary 1.2 to identify u as the unique variational solution.
  • standard math Amann-Escher differentiation theorem for series in V'
    Used in Lemma 4.1 to justify differentiating the uniformly convergent series sum v_j in C(R; V').
  • standard math Standard Sobolev, product-rule, and Bochner-space facts
    Used throughout, e.g., the product rule in the estimate of epsilon L_beta v and the boundedness of zero-extension maps in Section 6.
  • domain assumption a is real-valued, measurable, uniformly elliptic, with a-1 in C^{0,1/2}([0,1]; L^infty)
    This is the hypothesis class of Lions' problem; the paper constructs such an a, so it is verified rather than assumed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Lions' Maximal Regularity Problem for Divergence-Form Differential Operators: Failure at the $\frac{1}{2}$-H\"older Endpoint." pith.science (2026). https://pith.science/paper/V7UNHFWU

@misc{pith2026260811194,
  author       = {Pith},
  title        = {Pith review of: Lions' Maximal Regularity Problem for Divergence-Form Differential Operators: Failure at the $\frac12$-H\"older Endpoint},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/V7UNHFWU}},
  note         = {Machine review of arXiv:2608.11194}
}
abstract

In this work we give a counterexample to maximal $\mathrm{L}^2$-regularity in Lions' problem for divergence-form differential operators. On a bounded interval, we construct a bounded, uniformly elliptic, real scalar diffusion coefficient that is $\frac{1}{2}$-H\"older continuous in time with values in spatial $\mathrm{L}^\infty$. It can be chosen arbitrarily close to the constant coefficient of the heat equation. For zero initial data and a forcing term that is continuous in time with square-integrable spatial values, the unique Lions variational solution has a time derivative that is not square integrable in space-time. Thus $\frac{1}{2}$-H\"older continuity alone does not imply maximal $\mathrm{L}^2$-regularity, even for arbitrarily small scalar perturbations of the heat equation. The construction is based on a lacunary family of oscillatory trigonometric modes localised on shrinking time intervals. The spatial profile and the oscillatory modes, together with their first spatial derivatives, vanish at both endpoints. This permits zero extension of the counterexample to the real line. Tensorisation and localisation by parabolic rescaling then yield real symmetric isotropic counterexamples on $\mathbb{R}^d$ and on every bounded domain $\Omega\subset \mathbb{R}^d$, for all $d\ge1$.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

14 extracted references · 13 canonical work pages

  1. [1]

    Achache and E

    M. Achache and E. M. Ouhabaz,Lions’ maximal regularity problem withH 1/2-regularity in time, J. Differential Equations266(2019), no. 6, 3654–3678, doi:10.1016/j.jde.2018.09.015

    work page doi:10.1016/j.jde.2018.09.015 2019

  2. [2]

    Amann and J

    H. Amann and J. Escher,Analysis I, translated from the German by G. Brookfield, Birkh¨ auser, Basel, 2005, doi:10.1007/b137107

    work page doi:10.1007/b137107 2005

  3. [3]

    Arendt, D

    W. Arendt, D. Dier, and S. Fackler,J. L. Lions’ problem on maximal regularity, Arch. Math. (Basel)109(2017), no. 1, 59–72, doi:10.1007/s00013-017-1031-6

    work page doi:10.1007/s00013-017-1031-6 2017

  4. [4]

    Auscher and M

    P. Auscher and M. Egert,On non-autonomous maximal regularity for elliptic operators in divergence form, Arch. Math. (Basel)107(2016), no. 3, 271–284, doi:10.1007/s00013-016-0934-y

    work page doi:10.1007/s00013-016-0934-y 2016

  5. [5]

    Bechtel, C

    S. Bechtel, C. Mooney, and M. Veraar,Counterexamples to maximal regularity for operators in divergence form, Arch. Math. (Basel)123(2024), no. 2, 199–209, doi:10.1007/s00013-024-02014-9

    work page doi:10.1007/s00013-024-02014-9 2024

  6. [6]

    de Simon,Un ’applicazione della teoria degli integrali singolari allo studio delle equazioni differenziali lineari astratte del primo ordine, Rend

    L. de Simon,Un ’applicazione della teoria degli integrali singolari allo studio delle equazioni differenziali lineari astratte del primo ordine, Rend. Sem. Mat. Univ. Padova34(1964), 205–223, Numdam

    work page 1964

  7. [7]

    Dier,Non-autonomous Cauchy problems governed by forms: maximal regularity and invariance, Ph.D

    D. Dier,Non-autonomous Cauchy problems governed by forms: maximal regularity and invariance, Ph.D. thesis, Universit¨ at Ulm, 2014

    work page 2014

  8. [8]

    Dier,Non-autonomous maximal regularity for forms of bounded variation, J

    D. Dier,Non-autonomous maximal regularity for forms of bounded variation, J. Math. Anal. Appl. 425(2015), no. 1, 33–54, doi:10.1016/j.jmaa.2014.12.006

    work page doi:10.1016/j.jmaa.2014.12.006 2015

Show all 14 references
  1. [9]

    Dier and R

    D. Dier and R. Zacher,Non-autonomous maximal regularity in Hilbert spaces, J. Evol. Equ.17 (2017), no. 3, 883–907, doi:10.1007/s00028-016-0343-5

    2017 doi

  2. [10]

    Fackler,J.-L

    S. Fackler,J.-L. Lions’ problem concerning maximal regularity of equations governed by non-autonomous forms, Ann. Inst. H. Poincar´ e C Anal. Non Lin´ eaire34(2017), no. 3, 699–709, doi:10.1016/j.anihpc.2016.05.001

    2017 doi

  3. [11]

    B. H. Haak and E. M. Ouhabaz,Maximal regularity for non-autonomous evolution equations, Math. Ann.363(2015), no. 3–4, 1117–1145, doi:10.1007/s00208-015-1199-7. LIONS’ MAXIMAL REGULARITY PROBLEM FOR DIFFERENTIAL OPERATORS 21

    2015 doi

  4. [12]

    N. V. Krylov,On parabolic equations in one space dimension, Comm. Partial Differential Equations41(2016), no. 4, 644–664, doi:10.1080/03605302.2015.1126734

    2016

  5. [13]

    Lions, ´Equations diff´ erentielles op´ erationnelles et probl` emes aux limites, Grundlehren der mathematischen Wissenschaften, vol

    J.-L. Lions, ´Equations diff´ erentielles op´ erationnelles et probl` emes aux limites, Grundlehren der mathematischen Wissenschaften, vol. 111, Springer, Berlin–G¨ ottingen–Heidelberg, 1961, doi:10.1007/978-3-662-25839-2

    1961 doi

  6. [14]

    E. M. Ouhabaz and C. Spina,Maximal regularity for non-autonomous Schr¨ odinger type equations, J. Differential Equations248(2010), no. 7, 1668–1683, doi:10.1016/j.jde.2009.10.004. (Lukas Niebel)ETH Z ¨urich, Department of Mathematics, R¨amistrasse 101, 8092 Z¨urich, Switzerlan...

    2010 doi

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.