pith. sign in

arxiv: 2604.23161 · v1 · submitted 2026-04-25 · 🧮 math.FA

Nonexistence of Henkin type projections via a Wiener theorem for multipliers

Eduard Curc\u{a} , Micha{\l} Wojciechowski This is my paper

Pith reviewed 2026-05-08 07:04 UTC · model grok-4.3

classification 🧮 math.FA
keywords Fourier multipliersnoncomplemented subspacesWiener theoremHenkin projectionsSobolev spacescontinuous functionstorusmatrix differential operators
0
0 comments X

The pith

For d at least 2, certain matrix differential operators define noncomplemented A(D)-free subspaces in Sobolev and continuous function spaces on the torus.

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

The paper extends a 1967 result of Henkin to show that in spaces X such as W^{l,1}, W^{l,∞} or C^l on the d-dimensional torus with d ≥ 2, the subspace of X^N consisting of elements annihilated by appropriate N×N matrix operators A(D) is not complemented. The proof proceeds by first establishing a new property: the kernel of any Fourier multiplier bounded on X obeys a weakened form of Wiener's theorem on singularities of measures. A reader would care because the existence of a bounded projection onto a subspace determines whether one can continuously split the ambient space, and its absence reveals rigidity in the geometry of these function spaces.

Core claim

We extend a result of Henkin (1967), showing that, for appropriate N×N matrix operators A(D), the subspace of X^N consisting of A(D)-free elements is noncomplemented, where X is one of W^{l,1}(T^d), W^{l,∞}(T^d) or C^l(T^d) for d ≥ 2 and l ≥ 0. In order to prove this we establish a new property of the Fourier multipliers that are bounded on X: the kernel k of any such multiplier obeys a weaker version of Wiener's theorem for the singularities of measures.

What carries the argument

The weaker version of Wiener's theorem for singularities of measures satisfied by the kernel k of any Fourier multiplier bounded on X; this property is used to prove that no bounded projection onto the A(D)-free subspace can exist.

If this is right

  • The A(D)-free subspace admits no bounded projection inside X^N.
  • The noncomplementedness holds uniformly for the three families of spaces W^{l,1}, W^{l,∞} and C^l on the torus.
  • The same conclusion applies to any matrix operator A(D) satisfying the appropriate conditions used in the extension of Henkin's argument.
  • The new multiplier-kernel property is sufficient to rule out bounded projections in this setting.

Where Pith is reading between the lines

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

  • The same kernel property might be used to study projections or bases in related spaces such as L^p on the torus.
  • The result suggests that questions about complemented subspaces for differential operators on compact manifolds could be approached via multiplier kernels.
  • One could test whether the weaker Wiener-type condition holds for multipliers on other Banach spaces of periodic functions.

Load-bearing premise

The kernel of any Fourier multiplier bounded on X satisfies a weaker version of Wiener's theorem for the singularities of measures.

What would settle it

Explicit construction of a bounded linear projection from X^N onto the subspace of A(D)-free elements, for some d ≥ 2, l ≥ 0, X among the listed spaces, and suitable matrix operator A(D).

read the original abstract

Let $d\geq 2$, $l\geq 0$ and suppose $X$ is one of the function spaces $W^{l,1}(\mathbb{T}^{d})$, $W^{l,\infty }(\mathbb{T}^{d})$ or $C^{l}(\mathbb{T}^{d})$. We extend a result of Henkin (1967), showing that, for appropriate $N\times N$ matrix operators $A(D)$, the subspace of $X^{N}$ consisting of $A(D)-$free elements is noncomplemented. In order to prove this we establish a new property of the Fourier multipliers that are bounded on $X$: the kernel $k$ of any such multiplier obeys a weaker version of Wiener's theorem for the singularities of measures.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 3 minor

Summary. The paper extends Henkin's 1967 result to show that, for d ≥ 2, l ≥ 0 and X one of W^{l,1}(T^d), W^{l,∞}(T^d) or C^l(T^d), the subspace of X^N consisting of A(D)-free elements is noncomplemented for suitable N×N matrix Fourier multipliers A(D). The proof proceeds by first establishing (Section 3) a weaker Wiener-type property for the kernels of bounded multipliers on these spaces via direct Fourier-side estimates and singularity analysis of measures, then applying it (Section 4) to any hypothetical bounded projection by averaging to produce a translation-invariant operator whose kernel would violate the new property.

Significance. If the central claims hold, the work provides a technically solid extension of classical results on complemented subspaces in Sobolev and Hölder-type spaces on the torus, with the new multiplier property potentially useful in broader harmonic analysis contexts. The direct-estimate approach to the Wiener-type property and the translation-invariance reduction are strengths that make the argument self-contained within the manuscript's scope.

major comments (2)
  1. [§3] §3: The weaker Wiener property is stated for kernels of multipliers bounded on X; the singularity analysis must be verified to cover the precise class of measures arising from the Fourier multipliers on W^{l,1} and C^l (including the l=0 endpoint), as this property is load-bearing for the contradiction in §4.
  2. [§4] §4, averaging step: the reduction from an arbitrary projection to a translation-invariant one requires that the averaged operator remains a projection onto the A(D)-free subspace; the manuscript should explicitly confirm that the kernel of the averaged operator still satisfies the hypotheses needed to invoke the §3 property.
minor comments (3)
  1. [Introduction] The specific matrix operators A(D) used in the main theorem are described as 'appropriate' in the abstract and introduction; an explicit example or minimal set of conditions on A(D) should be given early to make the statement self-contained.
  2. Notation for the Fourier multipliers and the A(D)-free subspace is introduced without a dedicated preliminary section; a short notation table or paragraph would improve readability for readers outside the immediate subfield.
  3. [Introduction] The reference to Henkin (1967) is cited but the precise statement being extended is not restated; including a one-sentence recap of the original result would clarify the novelty.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading of the manuscript and the recommendation of minor revision. The two major comments identify points where the exposition can be strengthened for clarity, particularly regarding the scope of the singularity analysis and the details of the averaging argument. We address each below and will incorporate the suggested clarifications in the revised version.

read point-by-point responses

  1. Referee: §3: The weaker Wiener property is stated for kernels of multipliers bounded on X; the singularity analysis must be verified to cover the precise class of measures arising from the Fourier multipliers on W^{l,1} and C^l (including the l=0 endpoint), as this property is load-bearing for the contradiction in §4.

    Authors: We agree that explicit verification strengthens the argument. Section 3 derives the weaker Wiener property via direct Fourier coefficient estimates and singularity analysis that apply to any measure whose Fourier transform satisfies the decay bounds implied by boundedness of the multiplier on the given X. These bounds are uniform across the listed spaces and hold at the endpoint l=0 (where W^{0,1}=L^1 and C^0=C). The analysis does not rely on l>0 regularity beyond the multiplier boundedness assumption. In the revision we will insert a short paragraph after the main statement in §3 that explicitly checks the l=0 cases and confirms that the measures arising from multipliers on W^{l,1} and C^l fall within the class treated by the singularity estimates. revision: yes

  2. Referee: §4, averaging step: the reduction from an arbitrary projection to a translation-invariant one requires that the averaged operator remains a projection onto the A(D)-free subspace; the manuscript should explicitly confirm that the kernel of the averaged operator still satisfies the hypotheses needed to invoke the §3 property.

    Authors: The A(D)-free subspace is translation-invariant because A(D) is a Fourier multiplier. Consequently, if P is any bounded projection onto this subspace, its average P_avg over the torus is a translation-invariant bounded operator on X^N that is still a projection onto the same subspace. Since P_avg is translation-invariant it is a Fourier multiplier, and boundedness on X^N is preserved by averaging; hence the kernel of P_avg satisfies all hypotheses of the §3 property. In the revision we will add two sentences in §4 making this invariance and preservation explicit before invoking the contradiction from §3. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation is self-contained

full rationale

The manuscript proves a new weaker version of Wiener's theorem for the kernels of bounded Fourier multipliers on W^{l,1}, W^{l,∞} and C^l via direct estimates on the Fourier side and singularity analysis (Section 3). This independent property is then applied in Section 4 to specific matrix multipliers A(D) by averaging any hypothetical projection to obtain a translation-invariant operator whose kernel would violate the established property, yielding noncomplementedness. No load-bearing step reduces by definition, by fitting, or by self-citation chain to the target claim; the argument relies on explicit estimates and standard averaging arguments that are externally verifiable.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The argument rests on standard properties of the torus and the listed function spaces together with the newly proved multiplier property; no free parameters or invented entities are visible from the abstract.

axioms (1)
  • domain assumption The spaces X = W^{l,1}(T^d), W^{l,∞}(T^d), C^l(T^d) admit a theory of Fourier multipliers whose kernels satisfy a weakened Wiener condition on measure singularities.
    This is the new property introduced in the paper and invoked to prove noncomplementedness.

pith-pipeline@v0.9.0 · 5426 in / 1319 out tokens · 64366 ms · 2026-05-08T07:04:26.642015+00:00 · methodology

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

17 extracted references · 17 canonical work pages

  1. [1]

    In: Colloquium Publications, Vol

    Benyamini, Y., Lindenstrauss, J.,Geometric Nonlinear Functional Analysis: Volume 1. In: Colloquium Publications, Vol. 48, American Mathematical Society Colloquium Publications, Providence, 2000

    work page 2000

  2. [2]

    Bonami, A., Mohanty, P.,Examples of Fourier Multipliers of the Sobolev SpaceW 1,1(Rd), Analysis Mathematica, Volume 44 no. 3, 2018

    work page 2018

  3. [3]

    Bonami, A., Poornima, S., Nonmultipliers of the Sobolev spacesW k,1(Rd), J. Funct. Anal. 71 (1), 175–181, 1987

    work page 1987

  4. [4]

    Curc˘ a, E.,On the continuity of Fourier multipliers on ˙W l,1(Rd)and ˙W l,∞(Rd).Journal of Functional Analysis, page 109573, 2022

    work page 2022

  5. [5]

    Curc˘ a, E.,On the interpolation of the spacesW l,1(Rd)andW r,∞(Rd), Rev. Mat. Iberoam. 40, no. 3, 931-986, 2024

    work page 2024

  6. [6]

    7(4), 526–528, 1965

    Decell Jr., H.P.,An application of the Cayley–Hamilton theorem to generalized matrix inver- sion.SIAM Rev. 7(4), 526–528, 1965

    work page 1965

  7. [7]

    M.,The nonexistence of a uniform homeomorphism between spaces of smooth functions of one and ofnvariables (n≥2), Matem

    Henkin, G. M.,The nonexistence of a uniform homeomorphism between spaces of smooth functions of one and ofnvariables (n≥2), Matem. sb., 74, 595-607, 1967

    work page 1967

  8. [8]

    Cambridge University Press, Cambridge, third edition, 2004

    Katznelson, Y.,An introduction to harmonic analysis.Cambridge Mathematical Library. Cambridge University Press, Cambridge, third edition, 2004. 31

    work page 2004

  9. [9]

    Kazaniecki, K., Wojciechowski, M.,On the continuity of Fourier multipliers on the homoge- neous Sobolev spaces ˙W 1,1(Rd), Ann. Inst. Fourier 66, 2013

    work page 2013

  10. [10]

    J Fourier Anal Appl 31, 63, 2025

    Kazaniecki, K., Wojciechowski, M.On the Analytic Version of the Mityagin-De Leeuw-Mirkil Non-Inequality on Bi-Disc Via Rudin-Shapiro Polynomials. J Fourier Anal Appl 31, 63, 2025

    work page 2025

  11. [11]

    V.,Sobolev embedding operators and the nonisomorphism of certain Banach spaces, Funktsional

    Kislyakov, S. V.,Sobolev embedding operators and the nonisomorphism of certain Banach spaces, Funktsional. Anal. i Prilozhen. 9:4, 22–27, 1975 (Russian). English transl.: Funct. Anal. Appl. 9, 290–294, 1975

    work page 1975

  12. [12]

    de Leeuw, K., Mirkil, H.,A priori estimates for differential operators inL ∞ norm,Ill. J. Math. 8, 112–124, 1964

    work page 1964

  13. [13]

    Analysis at Urbana

    Pe lczy´ nski, A.,Boundedness of the canonical projection for Sobolev spaces generated by finite families of linear differential operators.In: Berkson E, Peck T, Uhl J, eds. Analysis at Urbana. London Mathematical Society Lecture Note Series. Cambridge University Press: 395-415, 1989

    work page 1989

  14. [14]

    Poornima, S.,Multipliers of the Sobolev spaces, J. Funct. Anal., 45, 1-27, 1982

    work page 1982

  15. [15]

    Wiener, N.,The Fourier Integral and Certain of Its Applications, Dover, New York, 1933

    work page 1933

  16. [16]

    Wojciechowski, M.,On the strong type multiplier norms of rational functions in several vari- ables, Ill. J. Math. 42(4), 582-600, 1998

    work page 1998

  17. [17]

    P.,Weakly Differentiable Functions, Graduate Texts in Mathematics, vol

    Ziemer, W. P.,Weakly Differentiable Functions, Graduate Texts in Mathematics, vol. 120, Springer-Verlag, New York, 1989. 32

    work page 1989