Pith. sign in

REVIEW 5 minor 32 references

Galerkin approximation of linear problems in Banach and Hilbert spaces

T0 review · 0 major / 5 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 Galerkin approximation converges for every choice of approximating subspaces exactly when the continuous sesquilinear form is essentially coercive and satisfies uniqueness.

desk verdict Theorem 5.2 (essential coercivity + uniqueness ⇔ universal Galerkin) is the real result; the Banach-space existence theorem is correct but leans on a deep cited equivalence. read the letter →

arxiv 1908.03326 v3 pith:P345WXVD submitted 2019-08-09 math.NA cs.NA

Wolfgang Arendt , Isabelle Chalendar (LAMA) , Robert Eymard (LAMA) This is my paper · ORCID
classification math.NAcs.NA MSC 65N3047A0747A5246B20
keywords GalerkinapproximationsesquilinearcoerciveformspropertiesinBanachspacesessentialcoercivityuniversalconvergenceBanach-Nečas-BabuškaconditionAubin-Nitscheestimatesaddlepointproblems
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 asks when a conforming Galerkin method for a linear problem $a(u,v)=\langle L,v\rangle$ is guaranteed to converge without the usual coercivity assumption. In Hilbert spaces it proves that convergence for every choice of approximating subspaces, called the universal Galerkin property, holds exactly when the form is essentially coercive and satisfies uniqueness. In Banach spaces it shows that, for a well-posed problem, some choice of approximating subspaces yields a convergent Galerkin approximation exactly when the trial space has a finite-dimensional Schauder decomposition. The paper also proves that Galerkin convergence, the discrete inf-sup (BNB) condition, and convergence for the adjoint problem are equivalent, and it generalizes the Aubin-Nitsche error estimate to nonsymmetric forms. If these characterizations are correct, a numerical analyst can decide from the operator alone whether a conforming Galerkin method must converge.

What carries the argument

The central objects are approximating sequences $(V_n)$ of finite-dimensional subspaces, together with two structural conditions. The first is the uniform Banach-Nečas-Babuška (BNB) inf-sup condition, which is shown to be equivalent to Galerkin convergence and to its adjoint analogue. The second is essential coercivity: a form $a$ is essentially coercive when every weakly null sequence $(u_n)$ with $a(u_n,u_n)\to 0$ is norm-convergent to $0$, a notion that Theorem 4.3 reduces to the algebraic statement that a finite-rank perturbation of $a$ is coercive. In the Banach setting, the decisive mechanism is a finite-dimensional Schauder decomposition of the trial space $U$, which is equivalent to the bounded projection approximation property in separable reflexive spaces and supplies the test subspaces for the Galerkin scheme.

What would settle it

Compute the discrete inf-sup constants for the non-essentially-coercive invertible form $a(u,v)=\sum_{n=0}^\infty (-1)^n u_n\overline{v_n}$ on $\ell^2$ using an approximating sequence that deliberately pairs coordinates $2n$ and $2n+1$, as in the paper's Example 4.9. Theorem 5.2 predicts these constants tend to zero; if some approximating sequence keeps them bounded away from zero, the Hilbert characterization is wrong. For the Banach statement, the decisive test is whether a separable reflexive Banach space with the bounded projection approximation property but no finite-dimensional Schauder decomposition exists; such a space would break the equivalence in Theorem 3.2.

Watch

Extended reading notes

Core claim

The paper's central claim is a two-way characterization. On a separable Hilbert space, a continuous sesquilinear form $a$ has the universal Galerkin property if and only if it is essentially coercive and satisfies uniqueness, where essential coercivity means that every weakly null sequence $(u_n)$ with $a(u_n,u_n)\to 0$ must converge to $0$ in norm; equivalently, by Theorem 4.3, some compact perturbation of the associated operator is coercive. On separable reflexive Banach spaces, for a well-posed problem there exists a choice of approximating sequences producing a convergent Galerkin approximation if and only if the trial space $U$ has a finite-dimensional Schauder decomposition. The paper further establishes that convergence of the Galerkin approximation, the uniform Banach-Nečas-Babuška condition, and convergence of the adjoint Galerkin approximation are equivalent conditions for given approximating sequences.

Load-bearing premise

The argument depends on an imported Banach-space theorem: in the separable reflexive spaces treated here, the bounded projection approximation property is equivalent to having a finite-dimensional Schauder decomposition; if that theorem failed, the characterization of which well-posed problems admit some convergent Galerkin approximation would collapse.

Editorial extensions

If this is right

  • Coercivity is unnecessary: in Hilbert spaces, any essentially coercive form satisfying uniqueness gives universal Galerkin convergence, with the error controlled by $\operatorname{dist}(u,V_n)$ as in Céa's lemma.
  • Galerkin convergence, the uniform BNB condition, and dual Galerkin convergence are equivalent; the best inf-sup constants of a form and its adjoint agree in Hilbert spaces.
  • In separable reflexive Banach spaces, a well-posed problem admits some convergent conforming Galerkin scheme exactly when the trial space has a finite-dimensional Schauder decomposition.
  • The Aubin-Nitsche trick applies to nonsymmetric forms and data in arbitrary spaces $X\hookrightarrow V'$, giving uniform error bounds of the form $\|u-u_n\|_H \le (M^2/\beta)\gamma_n(X)\gamma_n^*(H)\|L\|_X$.
  • The discrete inf-sup conditions for saddle-point problems are necessary as well as sufficient for convergence of mixed approximations.

Reading between the lines

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

  • Editorial inference: the Hilbert characterization turns a numerical question into finite-dimensional linear algebra, because one can certify convergence on every mesh by finding a finite-rank projection $P$ and a constant $\alpha>0$ such that $|a(u,u)|+\|Pu\|^2\ge\alpha\|u\|^2$; the paper does not discuss this certification algorithm.
  • Editorial inference: since essential coercivity is stable under compact perturbations and is an open property, small nonsymmetric perturbations of coercive problems automatically fall inside the universal Galerkin class, so stabilization strategies can be certified by compactness rather than by mesh-dependent constants.
  • Editorial inference: the Banach-space obstruction suggests that for a separable reflexive space lacking the approximation property, some well-posed linear problems admit no convergent conforming Galerkin approximation under any choice of subspaces; the paper points to known counterexamples but does not exhibit such a problem explicitly.
  • Editorial inference: since the uniform BNB condition and its adjoint analogue are equivalent, a numerical code that monitors one inf-sup constant can diagnose the other, which may give a practical stability test for Petrov-Galerkin or saddle-point discretizations by solving the adjoint system instead.
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

0 major / 5 minor

Summary. The paper studies conforming Galerkin (Petrov-Galerkin) approximation of the linear problem a(u,v)=⟨L,v⟩ on Banach and Hilbert spaces. Section 2 proves that convergence of the Galerkin approximation, the uniform Banach-Nečas-Babuška condition, the dual BNB condition, and convergence of the dual Galerkin approximation are all equivalent. Section 3 shows that, for a well-posed problem on separable reflexive Banach spaces, the existence of some approximating sequences for which Galerkin approximations converge is equivalent to the space U having the bounded projection approximation property, and, via a cited result of Casazza, equivalent to U having a finite-dimensional Schauder decomposition. Sections 4-5 introduce essential coercivity and prove the central Hilbert-space theorem: a continuous sesquilinear form has the universal Galerkin property if and only if it is essentially coercive and satisfies uniqueness. Section 6 generalizes the Aubin-Nitsche argument to non-symmetric forms with data in interpolation spaces, producing two-scale error estimates. Section 7 applies the theory to selfadjoint operators with compact resolvent and to finite element approximation of a non-coercive elliptic problem with convection and reaction terms. Section 8 gives a converse of Brezzi's theorem for saddle point problems.

Significance. The results, if correct, are significant: they convert classical sufficient conditions into exact characterizations and identify essential coercivity as the precise property underlying universal Galerkin convergence. The proofs are detailed and largely self-contained; the paper is honest about the one deep external input, the Casazza equivalence used in Theorem 3.2. The applications are concrete and include optimal error estimates and a converse of Brezzi's conditions. The manuscript also contains several useful byproducts, including an explicit constant in Céa-type estimates, a self-contained proof of the equivalence of BNB and dual BNB, and a careful treatment of the Aubin-Nitsche trick for non-selfadjoint forms. No fitted parameters or circular arguments appear: the characterizations are proved from stated hypotheses, and the numerical examples are illustrations rather than evidence.

minor comments (5)
  1. [Section 3, proof of Theorem 3.2, implication (i)⇒(ii)] The proof should read 'Let u∈U' and 'P_n u∈U_n'; as printed, 'u∈V' and 'P_n u∈V_n' are incompatible with the domain of a on U×V and with the goal of exhibiting projections on U.
  2. [Proposition 2.9, Eq. (2.12)] The right-hand side of (2.12) should be β||v||_V, not β||u||_U; the subsequent proof and the stated equivalence show this is a typographical error.
  3. [Section 7.1, Eq. (7.1)] The eigenvalue index is missing in the displayed estimate: it should read |λ_n|^{-1-s/2}, consistent with the preceding computation using γ_n(V_s) and γ_n(H).
  4. [Theorem 4.5, proof of (i)⇒(ii)] The sentence 'Let V1 = ker V' should be 'Let V1 = ker P', where P is the finite-rank orthogonal projection from Theorem 4.3(ii); the subsequent decomposition of V uses this kernel and the range of P.
  5. [Section 3, Theorem 3.2] The implication (ii)⇒(iii) is imported from [7, Theorem 6.4(3)], and Definition 3.1(c) alone does not provide the nestedness property (3.1) used in the proof of (iii)⇒(i). The authors should state the precise hypotheses of the cited theorem so the reader can verify that the match with Definition 3.1(c) is exact.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper proves its characterizations from stated hypotheses, and imports only external lemmas (Casazza, standard FEM estimates) that do not contain the target results.

full rationale

The derivation chain in this paper is self-contained against independent definitions. Theorem 5.2 is proved directly: essential coercivity is defined sequentially (Definition 4.2), Theorem 4.3 converts it to a compact-perturbation condition, and the two directions of the universal Galerkin characterization are argued by weak-compactness and by constructing a perturbed approximating sequence; no parameter is fitted and no quantity is renamed as a prediction. Theorem 3.2 is also proved in the text for (i)->(ii) and (iii)->(i), while the only unproved implication (ii)->(iii) is imported from Casazza's survey [7, Theorem 6.4(3)], an external source that the paper explicitly notes is open outside the reflexive setting, and the paper itself flags the dependence. The self-citations that occur ([1] for compact ellipticity in Remark 4.10, [2] for standard finite-element estimates) are ancillary and not load-bearing; neither supplies the main equivalence. There is no step where Eq. X equals Eq. Y by construction, no fitted parameter renamed as a prediction, and no self-citation chain that forces the result. The appropriate finding is therefore no significant circularity.

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

The central theorems rest on standard functional analysis, the imported Banach-space equivalence from Casazza for BPAP/FDD, and, in the finite element application, convex-domain regularity and interpolation estimates. There are no fitted parameters or invented entities; essential coercivity is a new definition but is proven equivalent to compact perturbation coercivity.

assumptions (5)
  • domain assumption U and V are separable reflexive Banach spaces, and dim U_n = dim V_n for all n.
    Assumed in Section 2 for the Petrov-Galerkin theory; reflexivity is used for weak compactness in Proposition 2.5 and for the BPAP/FDD equivalence.
  • standard math Standard functional analysis tools: Hahn-Banach, Banach-Steinhaus, Mazur's theorem, Fredholm alternative, closed graph theorem.
    Used throughout Sections 2-6; e.g., Proposition 2.8 uses Banach-Steinhaus, Theorem 3.2 uses Mazur, Corollary 4.6 uses Fredholm alternative.
  • standard math Casazza's theorem [7, Theorem 6.4(3)]: in separable reflexive spaces, BPAP is equivalent to having a finite-dimensional Schauder decomposition.
    Imported in Theorem 3.2 to connect assertions (ii) and (iii).
  • standard math Kato's lemma: for any projection Q on a Hilbert space, ||Q|| = ||I-Q||.
    Used in Proposition 2.6 to improve the error constant to M/beta.
  • domain assumption Kadlec's H2-regularity on convex domains and the finite element interpolation estimate (7.5).
    Used in Theorem 7.3 and Theorem 7.6 to obtain H2 regularity and the h and h^2 rates; these fail for nonconvex domains or non-quasi-uniform meshes.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Galerkin approximation of linear problems in Banach and Hilbert spaces." pith.science (2026). https://pith.science/paper/P345WXVD

@misc{pith2026190803326,
  author       = {Pith},
  title        = {Pith review of: Galerkin approximation of linear problems in Banach and Hilbert spaces},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P345WXVD}},
  note         = {Machine review of arXiv:1908.03326}
}
abstract

In this paper we study the conforming Galerkin approximation of the problem: find u $\in$ U such that a(u, v) = <L, v> for all v $\in$ V, where U and V are Hilbert or Banach spaces, a is a continuous bilinear or sesquilinear form and L $\in$ V' a given data. The approximate solution is sought in a finite dimensional subspace of U, and test functions are taken in a finite dimensional subspace of V. We provide a necessary and sufficient condition on the form a for convergence of the Galerkin approximation, which is also equivalent to convergence of the Galerkin approximation for the adjoint problem. We also characterize the fact that U has a finite dimensional Schauder decomposition in terms of properties related to the Galerkin approximation. In the case of Hilbert spaces, we prove that the only bilinear or sesquilinear forms for which any Galerkin approximation converges (this property is called the universal Galerkin property) are the essentially coercive forms. In this case, a generalization of the Aubin-Nitsche Theorem leads to optimal a priori estimates in terms of regularity properties of the right-hand side L, as shown by several applications. Finally, a section entitled "Supplement" provides some consequences of our results for the approximation of saddle point problems.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 31 canonical work pages

  1. [1]

    Arendt, A

    W. Arendt, A. F. M. ter Elst, J. B. Kennedy, and M. Sauter. The Dirichlet-to-Neumann operator via hidden compactness. J. Funct. Anal. , 266(3):1757–1786, 2014

    work page 2014

  2. [2]

    Arendt and K

    W. Arendt and K. Urban. Partielle Differenzialgleichungen. Eine Einf¨ uhrung in an alytische und numerische Methoden. Berlin: Springer Spektrum, 2nd edition edition, 2018

    work page 2018

  3. [3]

    Babuˇ ska

    I. Babuˇ ska. Error-bounds for finite element method. Numer. Math. , 16:322–333, 1970/71

    work page 1970

  4. [4]

    F. Brezzi. On the existence, uniqueness and approximation of sa ddle-point problems arising from Lagrangian multipliers. Rev. Fran¸ caise Automat. Informat. Recherche Op´ erationnelle S´ er. Rouge, 8(R-2):129–151, 1974

    work page 1974

  5. [5]

    Brezzi and M

    F. Brezzi and M. Fortin. Mixed and hybrid finite element methods , volume 15 of Springer Series in Computational Mathematics . Springer-Verlag, New York, 1991

    1991

  6. [6]

    F. E. Browder. Nonlinear operators and nonlinear equations of e volution in Banach spaces. In Non- linear functional analysis (Proc. Sympos. Pure Math., Vol. XVIII, Part 2, Chicago, Ill., 1968) , pages 1–308, 1976

    work page 1968

  7. [7]

    P. G. Casazza. Chapter 7 - approximation properties. In W. Joh nson and J. Lindenstrauss, editors, Handbook of the Geometry of Banach Spaces , volume 1 of Handbook of the Geometry of Banach Spaces, pages 271 – 316. Elsevier Science B.V., 2001

    work page 2001

  8. [8]

    Chesnel and P

    L. Chesnel and P. jun. Ciarlet. T -coercivity and continuous Galerkin methods: application to trans- mission problems with sign changing coefficients. Numer. Math. , 124(1):1–29, 2013

    work page 2013

Show all 32 references
  1. [9]

    S. H. Christiansen. Discrete Fredholm properties and converge nce estimates for the electric field integral equation. Math. Comput. , 73(245):143–167, 2004

    2004

  2. [10]

    Droniou, T

    J. Droniou, T. Gallou¨ et, and R. Herbin. A finite volume scheme fo r a noncoercive elliptic equation with measure data. SIAM J. Numer. Anal. , 41(6):1997–2031, 2003

    1997

  3. [11]

    P. Enflo. A counterexample to the approximation problem in Bana ch spaces. Acta Math. , 130:309– 317, 1973

    1973

  4. [12]

    Ern and J.-L

    A. Ern and J.-L. Guermond. Theory and Practice of Finite Elements , volume 159 of Applied Math- ematical Sciences. Springer-Verlag, New York, 2004

    2004

  5. [13]

    Grisvard

    P. Grisvard. Elliptic Problems in Nonsmooth Domains , volume 24 of Monographs and Studies in Mathematics. Pitman (Advanced Publishing Program), Boston, MA, 1985

    1985

  6. [14]

    Gustafson

    K. Gustafson. The Toeplitz–Hausdorff theorem for linear oper ators. Proc. Amer. Math. Soc., 25:203– 204, 1970

    1970

  7. [15]

    Hackbusch

    W. Hackbusch. Theorie und Numerik elliptischer Differentialgleichungen . Heidelberg: Springer Spek- trum, 4th revised edition edition, 2017

    2017

  8. [16]

    J. Kadlec. On the regularity of the solution of the Poisson proble m on a domain with boundary locally similar to the boundary of a convex open set. Czech. Math. J. , 14:386–393, 1964. 32 W. ARENDT, I. CHALENDAR, AND R. EYMARD

    1964

  9. [17]

    T. Kato. Estimation of iterated matrices, with application to the von Neumann condition. Numer. Math., 2:22–29, 1960

    1960

  10. [18]

    O. A. Ladyzhenskaya. The mathematical theory of viscous incompressible flow . Revised English edi- tion. Translated from the Russian by Richard A. Silverman. Gordon a nd Breach Science Publishers, New York-London, 1963

    1963

  11. [19]

    Le Bris, F

    C. Le Bris, F. Legoll, and F. Madiot. Stabilisation de probl` emes n on coercifs via une m´ ethode num´ erique utilisant la mesure invariante.C. R., Math., Acad. Sci. Paris , 354(8):799–803, 2016

    2016

  12. [20]

    Lindenstrauss and L

    J. Lindenstrauss and L. Tzafriri. Classical Banach spaces. I . Springer-Verlag, Berlin-New York, 1977. Sequence spaces, Ergebnisse der Mathematik und ihrer Grenzgeb iete, Vol. 92

    1977

  13. [21]

    Lions and E

    J.-L. Lions and E. Magenes. Probl` emes aux limites non homog` enes et applications. Vol . 1 . Travaux et Recherches Math´ ematiques, No. 17. Dunod, Paris, 1968

    1968

  14. [22]

    W. V. Petryshyn. On projectional-solvability and the Fredholm a lternative for equations involving linear A-proper operators. Arch. Rational Mech. Anal. , 30:270–284, 1968

    1968

  15. [23]

    W. V. Petryshyn. On the approximation-solvability of equations involving A-proper and psuedo- A- proper mappings. Bull. Amer. Math. Soc. , 81:223–312, 1975

    1975

  16. [24]

    A. Pietsch. History of Banach Spaces and Linear Operators . Birkh¨ auser, Basel, 2007

    2007

  17. [25]

    Pr¨ ossdorf and B

    S. Pr¨ ossdorf and B. Silbermann. Numerical analysis for integral and related operator equat ions, volume 52 of Operator Theory: Advances and Applications . Birkh¨ auser Verlag, Basel, 1991

    1991

  18. [26]

    C. J. Read. Different forms of the approximation property. Ty ped manuscript, Leeds, 1986

    1986

  19. [27]

    A. H. Schatz. An observation concerning Ritz-Galerkin method s with indefinite bilinear forms. Math- ematics of Computation , 28(128):959–962, 1974

    1974

  20. [28]

    A. H. Schatz and J. Wang. Some new error estimates for ritz-g alerkin methods with minimal regu- larity assumptions. Mathematics of Computation , 65(213):19–27, 1996

    1996

  21. [29]

    A. Stern. Banach space projections and Petrov-Galerkin est imates. Numer. Math. , 130(1):125–133, 2015

    2015

  22. [30]

    S. J. Szarek. A Banach space without a basis which has the boun ded approximation property. Acta Math., 159:81–98, 1987

    1987

  23. [31]

    Xu and L

    J. Xu and L. Zikatanov. Some observations on Babuˇ ska and Br ezzi theories. Numer. Math. , 94(1):195–202, 2003

    2003

  24. [32]

    E. Zeidler. Nonlinear functional analysis and its applications. II/A . Springer-Verlag, New York, 1990. Linear monotone operators, Translated from the German by the a uthor and Leo F. Boron. Wolfgang Arendt, Institute of Applied Analysis, Universit y of Ulm. Helmholtzstr. 18, ...

    1990

Pith tools

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