On generalized limits and ultrafilters

Cihan Orhan; Paolo Leonetti

On generalized limits and ultrafilters

T0 review · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Every generalized limit is a Choquet average of ultrafilter limits.

desk verdict Central representation results are new and mostly correct; the false equality in Theorem 1.3 is an application-level error that should be fixed but doesn't sink the paper. read the letter →

arxiv 2505.23263 v1 pith:4IAY5QZT submitted 2025-05-29 math.FA math.GN

Paolo Leonetti , Cihan Orhan This is my paper · ORCID

classification math.FAmath.GN MSC 47B6554D3540A3554A20

keywords generalizedlimitsidealsonomegapositivelinearfunctionalsl-infinityultrafiltersChoquetintegralStone-CechcompactificationSI-limitsI-convergence

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

The paper studies generalized limits on bounded sequences: positive normalized linear functionals on $\ell_\infty$ that agree with the ordinary limit on convergent sequences and vanish on the characteristic sequences of sets in an ideal $\mathcal{I}$. Its central claim is that every such functional is a Choquet average of ultrafilter limit functionals, equivalently that the family $\mathrm{SL}(\mathcal{I})$ is exactly the weak* closed convex hull of the ultrafilter limit functionals attached to $\mathcal{I}$. This reduces a large, seemingly unwieldy convex family to its extreme points. The same representation then yields a diameter dichotomy, a characterization of differences of generalized limits, and a proof that the sequences on which all generalized limits agree are exactly the bounded $\mathcal{I}$-convergent sequences.

What carries the argument

The carrying object is the Stone–Čech compactification $\beta\omega$, viewed as the space of ultrafilters on $\omega$, cut down to the compact subspace $\mathrm{Ult}(\mathcal{I})$ of free ultrafilters $F$ whose corresponding maximal ideal contains $\mathcal{I}$. Each such ultrafilter defines the functional $f_F(x) = F\!\lim x$, and the proof establishes that these are precisely the extreme points of $\mathrm{SL}(\mathcal{I})$. Krein–Milman then turns the extreme points into the whole weak* compact convex set, while the identification of $\ell_\infty'$ with finitely additive measures on $2^\omega$ converts the same extreme-point structure into the Choquet-integral representation over $\mathrm{Ult}(\mathcal{I})$.

What would settle it

Exhibit an ideal $\mathcal{I}$ and a functional $f \in \mathrm{SL}(\mathcal{I})$ whose associated finitely additive measure lies outside the weak* closed convex hull of the $\{0,1\}$-valued measures $\mu_F$ for $F \in \mathrm{Ult}(\mathcal{I})$; equivalently, find an extreme point of $\mathrm{SL}(\mathcal{I})$ that is not any $f_F$. A direct search would try to separate such an $f$ from every finite convex combination of $f_F$'s by a weak* continuous functional, contradicting the claimed representation.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is Theorem 2.2: for every ideal $\mathcal{I}$ on $\omega$, $\mathrm{SL}(\mathcal{I}) = \mathrm{co}(\{f_F : F \in \mathrm{Ult}(\mathcal{I})\})$, where $\mathrm{Ult}(\mathcal{I})$ is the compact space of free ultrafilters containing the dual filter of $\mathcal{I}$ and $f_F(x) = F\!\lim x$. Equivalently, Theorem 2.1 states that each $f \in \mathrm{SL}(\mathcal{I})$ has the form $f(x) = \int_{\mathrm{Ult}(\mathcal{I})} F\!\lim x \, d\rho(F)$ for a normalized capacity $\rho$. In other words, every positive normalized linear functional that extends the limit and kills $\mathcal{I}$ is a mixture of the maximal-ideal limit functionals indexed by ultrafilters. From this convex-hull description the paper derives the diameter and difference theorems, the meager-ideal strengthening, and the recovery of Freedman's characterization of the space of sequences with a single generalized limit.

Load-bearing premise

The load-bearing premise is an externally established claim, quoted from another paper, that the extreme points of $\mathrm{SL}(\mathcal{I})$ are exactly the functionals $f_F$ coming from free ultrafilters; if that claim ever failed, the convex-hull representation and the diameter and difference theorems built on it would not follow.

Editorial extensions

If this is right

Every $\mathrm{SI}$-limit is determined by its values on characteristic functions of subsets of $\omega$, and those values form a normalized capacity on $\mathrm{Ult}(\mathcal{I})$.
The diameter of $\mathrm{SL}(\mathcal{I})$ is $2$ if and only if $\mathcal{I}$ is not maximal, so non-maximal ideals admit two generalized limits that are as far apart as two norm-one functionals can be.
For meager ideals the diameter bound is witnessed at every point: each $f \in \mathrm{SL}(\mathcal{I})$ has a partner $g$ with $\|f-g\|=2$.
A bounded sequence has a single value under all $\mathrm{SI}$-limits exactly when it is $\mathcal{I}$-convergent, recovering Freedman's description of $V(\mathcal{I})$.
The distance from a bounded sequence to the space of bounded $\mathcal{I}$-convergent sequences equals half the spread between its $\mathcal{I}$-limit inferior and $\mathcal{I}$-limit superior.

Reading between the lines

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

Editorial inference: the same convex-hull description should also govern translation-invariant Banach limits, which form a face of $\mathrm{SL}(\mathrm{Fin})$; the paper does not address the invariant subfamily.
Editorial inference: the Choquet representation suggests a concrete approximation scheme — approximating a given generalized limit by finite convex combinations of ultrafilter limits and measuring the weak* approximation rate — which the paper does not pursue.
Editorial inference: the unique-decomposition dichotomy in Theorem 2.5 hints that the face lattice of $\mathrm{SL}(\mathcal{I})$ could support a canonical split of a generalized limit into a part supported on $\mathrm{Ult}(\mathcal{I})$ and a purely non-atomic part; exploring that split is beyond the paper's stated scope.

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.

Desk Editor's Note

read the letter

Central representation results are new and mostly correct; the false equality in Theorem 1.3 is an application-level error that should be fixed but doesn't sink the paper.

The paper's real contribution is Theorem 2.2: every SI-limit is a weak* limit of convex combinations of ultrafilter limit functionals, with the Choquet-integral form in Theorem 2.1 as the operative representation. That unifies a line of work on Banach limits, statistical limits, and density limits under one framework. The diameter dichotomy (Theorem 2.3), the meager strengthening (Theorem 2.4), and the difference characterization (Theorem 2.5) are clean and follow from the representation. The proofs are standard but well-executed. I also like the distance formula in Proposition 5.1.

The external extreme-point identification from [10, Claim 4] is genuinely load-bearing but not a real risk: the stress-test note is right that it can be verified directly, so the self-citation is a non-issue. No hidden circularity.

The main defect is Theorem 1.3. The equality c(I*) ∩ ℓ∞ = c(I) ∩ ℓ∞ is false for general ideals; I* convergence is strictly stronger than I-convergence. The reader's counterexample works. The theorem as stated claims equalities that overshoot. That said, the proof of V(I) = c(I) ∩ ℓ∞ does not rely on the false equality—it uses cluster points and the Choquet representation in the two directions—so the intended result survives. The statement should be revised: keep V(I) = c(I) ∩ ℓ∞, attribute the c(I*) version to Freedman separately, and note the second equality holds only for P-ideals.

Minor: the P-ideal decomposition c(I) = c + c00(I) is quoted rather than proven, which is fine. The paper is honest about what is recovered versus new.

Who should read this: anyone working on generalized limits, summability theory, or the structure of ℓ∞* . It deserves a serious referee. I'd send it out, with instructions to fix Theorem 1.3 and double-check the statement of Freedman's theorem.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the imported extreme-point identification from [10] is load-bearing but is an independent, checkable result; no step reduces to its own input.

full rationale

The paper's central derivation chain passes Theorem 2.1 through [10, Theorem 1.1] and Theorem 2.2 through [10, Claim 4], both from a paper with overlapping authorship. These citations are load-bearing: the Krein-Milman step in Theorem 2.2 needs the identification ext(T[SL(I)]) = {mu_F : F in Ult(I)}, and Theorem 2.1 needs a capacity representation for I-invariant finitely additive measures. However, this is not circularity. The cited claims are parameter-free theorems about finitely additive measures, not restatements of the target results, and the extreme-point identification is independently checkable: any positive normalized measure vanishing on I that is not 0-1 valued can be decomposed into two distinct such measures by conditioning on a set A with 0 < mu(A) < 1, so only the Dirac measures of ultrafilters in Ult(I) can be extreme. The textbook reference [1, p. 544] corroborates the same fact. No parameter is fitted and no quantity is defined in terms of another quantity it is then said to predict. The questionable equality c(I*) cap l_infty = c(I) cap l_infty in Theorem 1.3 is a correctness concern about a cited folklore fact, not a circular reduction, and it does not feed back into the representation theorems. Therefore the honest finding is no significant circularity.

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

The central claims rest on standard functional analysis results (Krein-Milman, Alaoglu, Choquet integration) and on several cited theorems from prior papers, including [10, Theorem 1.1 and Claim 4] (by overlapping authors) and [22, Corollary 2.10]. No free parameters are fitted; no new entities are postulated.

assumptions (5)

standard math Krein-Milman theorem and Alaoglu's theorem apply to the weak* compact convex set SL(I)
Used in proof of Theorem 2.2 to conclude SL(I) = co({f_F}) from the extreme points.
domain assumption [10, Theorem 1.1]: every finitely additive normalized measure on ω can be represented as a Choquet integral of Dirac measures over an appropriate compact space
Used in proof of Theorem 2.1; cited from a prior paper by Cerreia-Vioglio, Leonetti, Maccheroni, Marinacci.
domain assumption [10, Claim 4]: the extreme points of T[SL(I)] are {μ_F : F ∈ Ult(I)}
Used in proof of Theorem 2.2 to identify ext(SL(I)); this is a substantive external result taken from a paper with overlapping authorship.
domain assumption [22, Corollary 2.10]: if I is meager, then for every f ∈ SL(I) there exists A ∉ I with f(1_A) = 0
Used in proof of Theorem 2.4; this is an external result about conglomerated filters and statistical measures.
ad hoc to paper [24, Theorem 2.4]: c(I⋆) ∩ ℓ∞ = c(I) ∩ ℓ∞
This 'well known' equality is cited in the proof of Theorem 1.3. It is false in general (e.g., for non-P-ideals), so this axiom is ad hoc and incorrect.

how reviews work

0 comments

Cite this review

Pith. "Pith review of On generalized limits and ultrafilters." pith.science (2026). https://pith.science/paper/4IAY5QZT

@misc{pith2026250523263,
  author       = {Pith},
  title        = {Pith review of: On generalized limits and ultrafilters},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4IAY5QZT}},
  note         = {Machine review of arXiv:2505.23263}
}

abstract

Given an ideal $\mathcal{I}$ on $\omega$, we denote by $\mathrm{SL}(\mathcal{I})$ the family of positive normalized linear functionals on $\ell_\infty$ which assign value $0$ to all characteristic sequences of sets in $\mathcal{I}$. We show that every element of $\mathrm{SL}(\mathcal{I})$ is a Choquet average of certain ultrafilter limit functionals. Also, we prove that the diameter of $\mathrm{SL}(\mathcal{I})$ is $2$ if and only if $\mathcal{I}$ is not maximal, and that the latter claim can be considerably strengthened if $\mathcal{I}$ is meager. Lastly, we provide several applications: for instance, recovering a result of Freedman in [Bull. Lond. Math. Soc. 13 (1981), 224--228], we show that the family of bounded sequences for which all functionals in $\mathrm{SL}(\mathcal{I})$ assign the same value coincides with the closed vector space of bounded $\mathcal{I}$-convergent sequences.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

33 extracted references · 30 canonical work pages

[1]

C. D. Aliprantis and K. C. Border,Infinite dimensional analysis, third ed., Springer, Berlin, 2006, A hitchhiker’s guide

work page 2006
[2]

C. D. Aliprantis and O. Burkinshaw,Positive operators, Springer, Dordrecht, 2006, Reprint of the 1985 original

work page 2006
[3]

R. E. Atalla,On the multiplicative behavior of regular matrices, Proc. Amer. Math. Soc.26 (1970), 437–446

work page 1970
[4]

Balcerzak and P

M. Balcerzak and P. Leonetti,On the relationship between ideal cluster points and ideal limit points, Topology Appl.252 (2019), 178–190

work page 2019
[5]

Bartoszewicz, P

A. Bartoszewicz, P. Das, and S. Gł¸ ab,On matrix summability of spliced sequences and A-density of points, Linear Algebra Appl.487 (2015), 22–42

work page 2015
[6]

Bartoszewicz, S

A. Bartoszewicz, S. Głab, and A. Wachowicz,Remarks on ideal boundedness, convergence and variation of sequences, J. Math. Anal. Appl.375 (2011), no. 2, 431–435. On Generalized Limits and Ultrafilters 15

work page 2011
[7]

Bennett and N

G. Bennett and N. J. Kalton,Consistency theorems for almost convergence, Trans. Amer. Math. Soc. 198 (1974), 23–43

work page 1974
[8]

Benyamini and J

Y. Benyamini and J. Lindenstrauss, Geometric nonlinear functional analysis. Vol. 1 , American Mathematical Society Colloquium Publications, vol. 48, American Mathematical Society, Providence, RI, 2000

work page 2000

Show all 33 references

[9]

Bühler and D

T. Bühler and D. A. Salamon,Functional analysis, Graduate Studies in Mathematics, vol. 191, American Mathematical Society, Providence, RI, 2018

2018
[10]

Cerreia-Vioglio, P

S. Cerreia-Vioglio, P. Leonetti, F. Maccheroni, and M. Marinacci,Capacities and Choquet averages of ultrafilters, Proc. Amer. Math. Soc.152 (2024), no. 3, 1139–1151

2024
[11]

Connor,The statistical and strongp-Cesàro convergence of sequences, Analysis 8 (1988), no

J. Connor,The statistical and strongp-Cesàro convergence of sequences, Analysis 8 (1988), no. 1-2, 47–63

1988
[12]

, R-type summability methods, Cauchy criteria,P-sets and statistical convergence, Proc. Amer. Math. Soc.115 (1992), no. 2, 319–327

1992
[13]

Connor and J

J. Connor and J. Kline,On statistical limit points and the consistency of statistical conver- gence, J. Math. Anal. Appl.197 (1996), no. 2, 392–399

1996
[14]

Connor and F

J. Connor and F. Temizsu, On Banach space norms for I-bounded sequences and their quotient, Positivity 27 (2023), no. 3, Paper No. 45, 9

2023
[15]

Farah, Analytic quotients: theory of liftings for quotients over analytic ideals on the integers, Mem

I. Farah, Analytic quotients: theory of liftings for quotients over analytic ideals on the integers, Mem. Amer. Math. Soc.148 (2000), no. 702, xvi+177

2000
[16]

Filipów, A

R. Filipów, A. Kwela, and P. Leonetti,Borel complexity of sets of ideal limit points, preprint, last updated: Apr 18, 2025 (arXiv:2411.10866)

2025 arXiv
[17]

Filipów and J

R. Filipów and J. Tryba,Representation of ideal convergence as a union and intersection of matrix summability methods, J. Math. Anal. Appl.484 (2020), no. 2, 123760, 21

2020
[18]

A. R. Freedman,Generalized limits and sequence spaces, Bull. London Math. Soc.13 (1981), no. 3, 224–228

1981
[19]

Fridy,Statistical limit points, Proc

J.A. Fridy,Statistical limit points, Proc. Amer. Math. Soc.118 (1993), no. 4, 1187–1192

1993
[20]

Fridy and C

J.A. Fridy and C. Orhan,Statistical limit superior and limit inferior, Proc. Amer. Math. Soc. 125 (1997), no. 12, 3625–3631

1997
[21]

Henriksen, Multiplicative summability methods and the Stone-Čech compactification, Math

M. Henriksen, Multiplicative summability methods and the Stone-Čech compactification, Math. Z. 71 (1959), 427–435

1959
[22]

Kadets, D

V. Kadets, D. Seliutin, and J. Tryba, Conglomerated filters and statistical measures, J. Math. Anal. Appl.509 (2022), no. 1, Paper No. 125955, 17

2022
[23]

Kania,A letter concerning Leonetti’s paper ‘Continuous projections onto ideal convergent sequences’, Results Math.74 (2019), no

T. Kania,A letter concerning Leonetti’s paper ‘Continuous projections onto ideal convergent sequences’, Results Math.74 (2019), no. 1, Paper No. 12, 4

2019
[24]

Kostyrko, M

P. Kostyrko, M. Mačaj, T. Šalát, and M. Sleziak, I-convergence and extremal I-limit points, Math. Slovaca55 (2005), no. 4, 443–464

2005
[25]

Leonetti and M

P. Leonetti and M. Caprio,Turnpike in infinite dimension, Canad. Math. Bull.65 (2022), no. 2, 416–430

2022
[26]

Leonetti and F

P. Leonetti and F. Maccheroni,Characterizations of ideal cluster points, Analysis (Berlin) 39 (2019), no. 1, 19–26

2019
[27]

Letavaj, L

P. Letavaj, L. Mišík, and M. Sleziak,Extreme points of the set of density measures, J. Math. Anal. Appl. 423 (2015), no. 2, 1150–1165. 16 P. Leonetti and C. Orhan

2015
[28]

Orhan and M

C. Orhan and M. Gülfirat,Generalized limits and ideal convergence, Filomat 38 (2025), no. 32, 11259—-11267

2025
[29]

Özgüç, E

İ. Özgüç, E. Taş, and T. Yurdakadim,Generalized limits and sequence of matrices, Positivity 24 (2020), no. 3, 553–563

2020
[30]

Rainwater, Regular matrices with nowhere dense support, Proc

J. Rainwater, Regular matrices with nowhere dense support, Proc. Amer. Math. Soc. 29 (1971), 361

1971
[31]

E. M. Semenov and F. A. Sukochev,Invariant Banach limits and applications, J. Funct. Anal. 259 (2010), no. 6, 1517–1541

2010
[32]

Tryba, Connections between matrix summability ideals and nonpathological analytic P- ideals, Fund

J. Tryba, Connections between matrix summability ideals and nonpathological analytic P- ideals, Fund. Math.265 (2024), no. 1, 1–13

2024
[33]

Yurdakadim, M

T. Yurdakadim, M. K. Khan, H. I. Miller, and C. Orhan,Generalized limits and statistical convergence, Mediterr. J. Math.13 (2016), no. 3, 1135–1149. Department of Economics, Università degli Studi dell’Insubria, via Monte Generoso 71, 21100 V arese, Italy Email address: leonet...

2016

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

C. D. Aliprantis and K. C. Border,Infinite dimensional analysis, third ed., Springer, Berlin, 2006, A hitchhiker’s guide

work page 2006

[2] [2]

C. D. Aliprantis and O. Burkinshaw,Positive operators, Springer, Dordrecht, 2006, Reprint of the 1985 original

work page 2006

[3] [3]

R. E. Atalla,On the multiplicative behavior of regular matrices, Proc. Amer. Math. Soc.26 (1970), 437–446

work page 1970

[4] [4]

Balcerzak and P

M. Balcerzak and P. Leonetti,On the relationship between ideal cluster points and ideal limit points, Topology Appl.252 (2019), 178–190

work page 2019

[5] [5]

Bartoszewicz, P

A. Bartoszewicz, P. Das, and S. Gł¸ ab,On matrix summability of spliced sequences and A-density of points, Linear Algebra Appl.487 (2015), 22–42

work page 2015

[6] [6]

Bartoszewicz, S

A. Bartoszewicz, S. Głab, and A. Wachowicz,Remarks on ideal boundedness, convergence and variation of sequences, J. Math. Anal. Appl.375 (2011), no. 2, 431–435. On Generalized Limits and Ultrafilters 15

work page 2011

[7] [7]

Bennett and N

G. Bennett and N. J. Kalton,Consistency theorems for almost convergence, Trans. Amer. Math. Soc. 198 (1974), 23–43

work page 1974

[8] [8]

Benyamini and J

Y. Benyamini and J. Lindenstrauss, Geometric nonlinear functional analysis. Vol. 1 , American Mathematical Society Colloquium Publications, vol. 48, American Mathematical Society, Providence, RI, 2000

work page 2000