REVIEW 1 major objections 4 minor 23 references
A study of weak$^*$-weak points of continuity in the unit ball of dual spaces
T0 review · 1 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash
Pith's one-line read The paper proves that on the unit ball of the second dual of a non-reflexive M-embedded Banach space, and consequently on unit balls of certain von Neumann algebras, the identity map has no point where weak* continuity implies weak…
desk verdict Solid M-ideal/predual paper, but the headline von Neumann algebra claim is false as stated — add 'non-reflexive' and everything holds together. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing mechanism is the correspondence, Lemma 3.1, that a functional $x^*\in S(X^*)$ is a point of $w^*$-$w$ continuity of the identity map $\mathrm{id}\colon(B(X^*),w^*)\to(B(X^*),w)$ exactly when it has a unique norm-preserving extension to $X^{**}$. An M-embedded space is one that is an M-ideal in its bidual, equivalently $X^{***}=X^*\oplus_1 X^\perp$; this decomposition supplies the L-projection used to push continuity up to third duals and the alternative-predual construction of Proposition III.2.10 used to destroy it. The construction chooses a norm-one functional on $X^{**}/X$, forms a $w^*$-closed hyperplane in $X^\perp$, and produces a new predual whose intersection with $X$ misses a prescribed unit vector. The positive density arguments write $S(X^*)$ as $B(X^*)\cap\bigcap_n O_n$ with $O_n=X^*\setminus(1-1/n)B(X^*)$, making the good points a $G_\delta$ and then transferring density to $S(X^{***})$.
What would settle it
The abstract's unconditional claim is already falsified by $M_n$: on a finite-dimensional unit ball the weak and weak* topologies coincide, so every point is a continuity point. For the intended theorem, the decisive test is to find a non-reflexive M-embedded space $X$ and a point $x^{**}\in B(X^{**})$ at which every $w^*$-convergent net is weakly convergent; Theorem 6.7 says no such point exists, so one example would settle the question.
Extended reading notes
Core claim
The central claim is Theorem 6.7: if $X$ is a non-reflexive M-embedded space, there is another predual $Y$ of $X^{**}$ with $Y\setminus X\neq\emptyset$, and the alternative predual can be chosen to avoid any prescribed point, so no element of $B(X^{**})$ is a point of continuity of $\mathrm{id}\colon(B(X^{**}),w^*)\to(B(X^{**}),w)$. The proof uses the failure of $X$ to be a strongly unique predual of $X^*$, as given in the cited monograph's Proposition III.2.10, and builds a $w^*$-closed hyperplane in $X^{***}$ that separates any candidate point from $X$. For von Neumann algebras, the structural decomposition $\mathcal{A}=\left(\bigoplus_{c_0}K(H_i)\right)^{**}$ for a predual with the Radon–Nikodym property makes $\mathcal{A}$ the second dual of a non-reflexive M-embedded space, so the no-point conclusion transfers. The paper also proves positive density results: for non-reflexive Hahn–Banach smooth spaces, $S(X^*)$ is a $w^*$-dense $G_\delta$ subset of $S(X^{***})$, and these are exactly the extreme-point continuity points when the dual is strictly convex. The statements should be read in the non-reflexive case; the abstract's unconditional formulation would fail for finite-dimensional algebras such as $M_n$.
Load-bearing premise
The von Neumann algebra conclusion rests on the structural decomposition of an RNP-predual von Neumann algebra as an $\ell_\infty$-sum of $L(H_i)$ and on the associated $c_0$-sum of compact operators being non-reflexive and M-embedded; if that sum is reflexive, as for finite-dimensional algebras, the conclusion fails.
Editorial extensions
If this is right
- For every non-reflexive M-embedded $X$, the identity map on $B(X^{**})$ is nowhere $w^*$-$w$ continuous: the weak and weak* topologies differ at every point of the dual unit ball.
- A non-reflexive von Neumann algebra whose predual has the Radon–Nikodym property has no $w^*$-$w$ continuity point on its unit ball, covering in particular infinite $\ell_\infty$-sums of full operator algebras $L(H_i)$.
- For non-reflexive Hahn–Banach smooth spaces, the continuity points are topologically large: $S(X^*)$ is a $w^*$-dense $G_\delta$ subset of $S(X^{***})$, and it coincides with the extreme-point continuity points when $X^*$ is strictly convex.
- The property does not travel stably between duals: renorming can create a continuity point in $B(X^*)$ that disappears in $B(X^{***})$, and $L^1[0,1]$ has no such point in its second dual but has an extreme one in its fourth dual.
- When $K(X)$ is an M-ideal in $L(X)$ and $X$ is non-reflexive, the points of $w^*$-norm continuity in $B(X^{***})$ form a $w^*$-dense $G_\delta$ subset of $S(X^{***})$.
Reading between the lines
- If the non-reflexivity hypothesis is made explicit, the von Neumann theorem suggests a dichotomy: reflexive algebras such as finite-dimensional sums of matrix algebras have all points continuous, while genuinely infinite ones have none; a complete classification by reflexivity of the associated $c_0$-sum is a natural next step.
- Via Lemma 3.1, the no-point conclusion is equivalent to every norm-one functional on $X$ having at least two norm-preserving extensions to $X^{**}$; this rephrasing could be tested by explicit extension arguments in concrete algebras such as $L(H)$.
- The fourth-dual resurrection for $L^1[0,1]$ hints that higher duals can restore continuity even when the second dual has none; one testable conjecture is that for non-atomic $L^1(\mu)$ spaces the even duals alternate between empty and non-empty sets of continuity points.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies the set of points where the identity map on the dual unit ball is continuous from the weak* topology to the weak topology, with emphasis on M-embedded spaces, higher duals, and von Neumann algebras. The main structural results are: for a non-reflexive M-embedded space X, B(X**) admits no point of weak*-weak continuity (Theorem 6.7); every point of S(X*) is such a point in B(X***) when X is M-embedded (Theorem 6.2); a renorming construction produces points of continuity that fail in the third dual (Theorem 7.3); and L1[0,1] has no such points in its second dual but has some in its fourth dual (Theorem 8.2). Section 9 contains Gδ-density results. The abstract and Proposition 6.8 claim that a von Neumann algebra whose predual has the Radon–Nikodym property has no point of weak*-weak continuity on its unit ball; this is false without a non-reflexivity hypothesis.
Significance. If the missing hypothesis is added, the paper is a solid contribution: the arguments are largely coherent, they rely on standard structural theorems such as M-ideal theory, Godefroy's lemma, and renorming results rather than on circular reasoning, and the corrected von Neumann algebra theorem gives a clean application of Theorem 6.7. The paper also contains useful concrete examples, including the fourth-dual phenomenon for L1[0,1]. However, the main advertised theorem is currently over-stated: finite-dimensional von Neumann algebras are counterexamples to the statement as written. The fix is local, but because the false statement appears in the abstract and in the central application, the paper needs revision before it can be accepted.
major comments (1)
- [Abstract and Proposition 6.8] The headline claim stated in the abstract and repeated in Proposition 6.8 is false as written. Take A = M_n. Its predual is the finite-dimensional space of trace-class operators on C^n, which has the Radon–Nikodym property; A is finite-dimensional, so the weak* and weak topologies coincide and every point of the unit ball is a point of continuity of the identity, directly contradicting the asserted nonexistence. In the proof of Proposition 6.8, the classification [HWW93, Proposition IV.2.9] yields A = (⊕_{c0} K(H_i))**, but Theorem 6.7 requires the M-embedded space to be non-reflexive, and for a finite family of finite-dimensional H_i the c0-sum is reflexive; the argument never checks this hypothesis, and the check is precisely what fails for M_n. The abstract and Proposition 6.8 should be amended by adding the hypothesis that A is non-reflexive (equivalently, infinite-dimensional), and the proof should note that non-reflexivity of A rules out the finite-dimensional case, so that ⊕_{c0} K(H_i) is non-reflexive and Theorem 6.7 applies.
minor comments (4)
- [Section 9, before Proposition 9.1] The sentence 'S(X*) is w*-dense in (B(X***), w*)' is not correct as written; Goldstine's theorem gives w*-density of B(X*) in B(X***). The density of S(X*) in S(X***) that Proposition 9.2 needs follows by a normalization argument and should be stated separately.
- [Lemma 2.3] The proof of Lemma 2.3 is omitted; since the passage from the unit ball to the unit sphere requires a normalization argument using the w*-lower semicontinuity of the norm to get ||y_α||→1, a sentence or two would improve readability.
- [Theorem 7.3] The last step of the proof is compressed: the assertion that two different norm-preserving extensions of f0 produce a failure of continuity for f in B(Z***) should be spelled out via Lemma 3.1 applied to Z**, as this is the key point of the construction.
- [General] There are minor typographical errors, such as 'focuss' in Section 3 and 'SP ACES' in the title, that should be corrected.
Circularity Check
No significant circularity: the central von Neumann algebra theorem rests on external structural theorems whose assumptions are independent of the target conclusion.
full rationale
The paper's derivation chain for the main von Neumann algebra result is self-contained relative to published external results. Proposition 6.8 reduces the von Neumann algebra case to Theorem 6.7 via the classification [HWW93, Proposition IV.2.9] and the M-embeddedness of the c0-sum of spaces of compact operators [HWW93, Example III.1.4]. Theorem 6.7 in turn relies on [HWW93, Proposition III.2.10], an external theorem asserting that a nonreflexive M-embedded space is not a strongly unique predual, together with a direct argument using Lemma 5.1 and a Hahn--Banach separation observation. None of these inputs is fitted to the paper's target claim, none is defined in terms of weak*-weak continuity, and none is a self-citation by the present authors; the cited external theorems have their own hypotheses and proofs that do not presuppose the paper's conclusions. The paper's self-citations ([BR98], [CGM24], [GMZ]) appear only in incidental remarks and examples, not as load-bearing premises of the main theorems. The abstract and Proposition 6.8 omit a nonreflexivity hypothesis, so they fail for finite-dimensional von Neumann algebras such as M_n, where weak* and weak topologies coincide; this is a genuine correctness defect, but it is a missing hypothesis, not a circular reduction. Consequently, no circular step is present, and the appropriate circularity score is 0.
Assumptions & free parameters
assumptions (5)
- standard math Standard functional analysis background: Hahn-Banach, Goldstine's theorem, weak*-compactness of dual unit balls, James' reflexivity theorem.
- domain assumption Godefroy's dictionary [Go81]: x* in S(X*) is a point of weak*-weak continuity iff it has a unique norm-preserving extension to X**.
- domain assumption M-ideal theory facts from HWW93: M-embedded spaces satisfy X*** = X* plus l1 X-perp, and non-reflexive M-embedded spaces fail to be strongly unique preduals of X*.
- domain assumption Decomposition theorem [HWW93, Proposition IV.2.9]: a von Neumann algebra whose predual has RNP is an ell-infinity sum of L(H_i), so A = (direct sum over c0 of K(H_i))**.
- domain assumption Renorming existence from [GI02, Proposition 4.1] and [Na06, Theorem 2.2] for very smooth points and Chebyshev non-strongly-proximinal subspaces.
Cite this review
Pith. "Pith review of A study of weak$^*$-weak points of continuity in the unit ball of dual spaces." pith.science (2026). https://pith.science/paper/TBEZT6NQ
@misc{pith2026250608458,
author = {Pith},
title = {Pith review of: A study of weak$^*$-weak points of continuity in the unit ball of dual spaces},
year = {2026},
howpublished = {\url{https://pith.science/paper/TBEZT6NQ}},
note = {Machine review of arXiv:2506.08458}
}
abstract
We study classes of Banach spaces where the points of weak$^*$-weak continuity for the identity mapping on the dual unit ball form a weak$^*$-dense and weak$^*$-$G_{\delta}$ set. We also discuss how this property behaves in higher duals of Banach spaces. We prove in particular that if $\mathcal{A}$ is a von Neumann algebra and its predual has the Radon--Nikod\'ym property, then there is no point of weak$^*$-weak continuity on the unit ball of $\mathcal{A}$.
Reference graph
Works this paper leans on
-
[1]
Basu, S.; Rao, T. S. S. R. K.: Some stability results for asymptotic norming properties in Banach spaces , Colloq. Math. 75 (1998) 271--284
work page 1998
-
[2]
D.: Geometric aspects of convex sets with the Radon-Nikod\'ym property
Bourgin, R. D.: Geometric aspects of convex sets with the Radon-Nikod\'ym property. Lecture Notes in Mathematics, 993. Springer-Verlag, Berlin, 1983. xii + 474 pp. ISBN: 3-540-12296-6
work page 1983
-
[3]
Civin, P., and Yood, B.: Quasi-reflexive spaces , Proc. Amer. Math. Soc. 8 (1957) 906--911
work page 1957
-
[4]
J.; Montesinos, V.: Some remarks on Phelps property U of a Banach space into C(K) spaces
Cobollo, C.; Guirao, A. J.; Montesinos, V.: Some remarks on Phelps property U of a Banach space into C(K) spaces . Rev. R. Acad. Cienc. Exactas Fís. Nat. Ser. A Mat. RACSAM 118 (2024), no. 1, Paper No. 6, 23 pp
work page 2024
-
[5]
Deville, R. ; Godefroy, G. ; Zizler, V.: Smoothness and renormings in Banach spaces , Pitman Monographs and Surveys in Pure and Applied Mathematics, vol. 64, Longman Scientific and Technical, Harlow, 1993
work page 1993
-
[6]
Diestel, J.: Geometry of Banach Spaces: Selected Topics , Lecture Notes in Mathematics, 485 , Springer, 1975
work page 1975
-
[7]
Diestel, J.; Uhl, J. J., Jr.: Vector measures . With a foreword by B. J. Pettis. Mathematical Surveys, No. 15. American Mathematical Society, Providence, RI, 1977. xiii +322 pp
work page 1977
-
[8]
Fabian, M., Habala, P., H\'ajek, P., Montesinos, V., and Zizler, V.: Banach spaces Theory: The Basis for Linear and Nonlinear Analysis , CMS Books in Mathematics, Springer, 2011
work page 2011
Show all 23 references
-
[9]
Israel J
Godefroy, G.: Points de Namioka, espaces normants, applications \`a la th\'eorie isometrique de la dualit\'e . Israel J. Math. 38 , 3 (1981), 209--220
1981
-
[10]
Godefroy, G.: Existence and uniqueness of isometric preduals: A survey. In B.-L. Lin, editor, Banach space theory (Proc., Iowa City, 1987), volume 85 of Contemp. Math., pages 131–193, Providence, RI, 1989. AMS
1987
-
[11]
Godefroy, G.; Indumathi, V.: Strong proximinality and polyhedral spaces . Rev. Mat. Complut. 14 (2001), no. 1, 105--125
2001
-
[12]
Set-Valued Anal
Godefroy, G.; Indumathi, V.: Norm-to-weak upper semi-continuity of the duality and pre-duality mappings . Set-Valued Anal. 10 (2002), no. 4, 317--330
2002
-
[13]
J.; Montesinos, V.; Zizler, V.: On preserved and unpreserved extreme points
Guirao, A. J.; Montesinos, V.; Zizler, V.: On preserved and unpreserved extreme points . In Descriptive Topology and Functional Analysis. In Honour of Jerzy Kakol's 60th Birthday. Springer Proceedings in Mathematics and Statistics Vol. 80, Springer, 2014
2014
-
[14]
Lecture Notes in Mathematics, 1547
Harmand, P.; Werner, D.; Werner, W.: M -ideals in Banach spaces and Banach algebras . Lecture Notes in Mathematics, 1547. Springer-Verlag, Berlin, 1993. viii +387 pp. ISBN: 3-540-56814-X
1993
-
[15]
B.: Geometric functional analysis and its applications
Holmes, R. B.: Geometric functional analysis and its applications . Graduate Texts in Mathematics, No. 24. Springer-Verlag, New York-Heidelberg, 1975. x +246 pp
1975
-
[16]
Illinois J
Hu, Z.; Lin, B-L.: RNP and CPCP in Lebesgue-Bochner function spaces . Illinois J. Math. 37 (1993), no. 2, 329--347
1993
-
[17]
E.: The isometric theory of classical Banach spaces
Lacey, H. E.: The isometric theory of classical Banach spaces . Die Grundlehren der mathematischen Wissenschaften, Band 208. Springer-Verlag, New York-Heidelberg, 1974. x +270 pp
1974
-
[18]
L.: Characterizations of denting points
Lin, B-L.; Lin, P-K.; Troyanski, S. L.: Characterizations of denting points . Proc. Amer. Math. Soc. 102 (1988). no.3, 526--528
1988
-
[19]
Lindenstrauss, J.; Tzafriri, L.: Classical Banach spaces. I. Sequence spaces . Ergebnisse der Mathematik und ihrer Grenzgebiete [Results in Mathematics and Related Areas], Band 92. Springer-Verlag, Berlin-New York, 1977. xiii +188 pp. ISBN: 3-540-08072-4
1977
-
[20]
Narayana, D.: Strong proximinality and renormings . Proc. Amer. Math. Soc. 134 (2006), no. 4, 1167--1172
2006
-
[21]
R.: Uniqueness of Hahn--Banach extensions and unique best approximation
Phelps, R. R.: Uniqueness of Hahn--Banach extensions and unique best approximation . Trans. Amer. Math. Soc. 95 (1960), 238--255
1960
-
[22]
Israel J
Raja, M.: On dual locally uniformly rotund norms . Israel J. Math. 29 , (2002), 77--91
2002
-
[23]
Studia Math
Valdivia, M.: On a class of Banach spaces . Studia Math. 60 (1977), no. 1, 11--13
1977
Reviewed August 7, 2026 · model on record in the stance chip above.
Discussion (0). Sign in to comment.