A Weak Topology on Metric Spaces

arXiv:2609.19368 · math.MG, math.FA, math.GN · Submitted 2026-09-16 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.

Ada: We're looking at the paper: "A Weak Topology on Metric Spaces".

Emmy: Here's the gist of the paper. A WEAK TOPOLOGY ON METRIC SPACES uses metric functionals to construct a weak topology associated with d-weak convergence on metric spaces and demonstrates that this topology is connected to weak metrics The primary result establishes the existence of a topology on a metric space induced by the space of all metric functionals, denoted as σ(X, X♢) This topology has three key properties: it is coarser than the metric topology on X, it makes every metric functional lower semi-continuous, and a net converges in this topology if and only if it converges d-weakly.

Ada: Construction of the Weak Topology The topology σ(X, X♢) is generated by a basis B defined using metric functionals Specifically, the basis element B(x, F, c) is defined as the set of points y such that h(y) > h(x) - c for all h in F.

Emmy: This construction ensures that a subset A is open if every point x in A has a basis element B in B such that x ∈ B ⊂ A.

Ada: The choice of base-point is irrelevant, as σ(X, X♢) equals σ(X, X♢p) when d(o, p) > 0.

Paper summary: Ada: So, let's go over what this paper is actually claiming about "A Weak Topology on Metric Spaces." It starts by establishing that they can use metric functionals to construct a weak topology that is associated with d-weak convergence on any metric space.

Emmy: They are essentially taking these metric functionals, which they call the space X♢, and using them to define a new notion of closeness between points in the space. This closeness defines the weak topology sigma(X, X♢).

Ada: The core thesis is that this topology has three properties: it’s coarser than the original metric topology on the space X, it makes every metric functional lower semi-continuous, and convergence in this new topology is exactly equivalent to d-weak convergence.

Kolya: So if you have a net converging in this new topology, it means that for every single functional h in that space X♢, the limit inferior of h evaluated along the net stays greater than or equal to the value of h at the limit point.

Emmy: That’s right. And they build a specific basis element B(x, F, c) using a finite set of functionals and a positive real number. This construction is what allows them to define what it means for an open set to contain a point in this weak topology around x.

Ada: The paper also shows that there’s an explicit link here with weak metrics. They prove the existence of a family of weak metrics, one for every functional in X♢, such that convergence in this new topology is equivalent to those individual weak metrics going to zero for all functionals.

Kolya: So, when we talk about d-weak convergence as the original notion, this paper provides a concrete way to check it by looking at these specific metrics delta h, where delta h(x, y) is defined based on the functional h.

Emmy: And they confirm that this connection holds in a very structured way. They show that for every point x and any other point y, the difference between the distance between them and their distances from a base point o relates to these functionals in a specific way.

Ada: It really lays out how you can use metric functionals as tools to probe convergence in metric spaces, moving beyond just standard metric notions.

Emmy: And they also make sure that for normed linear spaces, this setup works very well structurally, specifically showing the topology is Hausdorff in those cases and confirming that bounded sequences converge d-weakly if and only if they converge in the classical weak sense.

Conclusion: Ada: To wrap up on "A Weak Topology on Metric Spaces," we’ve seen how Gutiérrez and Nevanlinna use metric functionals to build a topology intrinsically linked to d-weak convergence.

Emmy: They successfully establish that this new topology has the claimed properties, especially its connection to the family of weak metrics they defined, which provides a very concrete way to characterize convergence.

Ada: It really shows that we can take an existing notion of convergence like d-weak convergence and build a topological space around it using tools from metric geometry.

Emmy: And the implication is that for certain spaces, like normed linear spaces, this topology is well-behaved; it’s Hausdorff and gives us a clean way to study these kinds of nets.

Ada: The paper also serves as an important correction to earlier results because it points out an error concerning the boundedness of d-weakly convergent sequences in normed linear spaces, which was shown by that specific sequence in Czero one <ref:2609.19368#pg1>.

Emmy: That correction is a key part of the contribution; it shows that we need to be careful about assuming boundedness just because you have d-weak convergence in these contexts.

Ada: So, essentially, they give us a robust topological structure tied to functional analysis and metric spaces that helps clarify what d-weak convergence actually means in practice.

Emmy: It gives us a new lens through which to view the relationship between algebraic structures, like metric functionals, and the analytical behavior of sequences.

math.MG, math.FA, math.GN

Submitted: 2026-09-16

Updated: 2026-11-10

Comments: Revised version with new results on the separation properties of the metric-functional weak topology and its comparison with the classical weak topology on normed linear spaces. The presentation has also been improved

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 65/100

The gist: A WEAK TOPOLOGY ON METRIC SPACES uses metric functionals to construct a weak topology associated with d-weak convergence on metric spaces and demonstrates that this topology is connected to weak…

Key concepts

Metric Functionals
These are functions defined on a metric space that take a point and return a real number. The paper uses these functionals to define the new topology, essentially using them as tools to measure convergence in a way related to the underlying metric.
d-Weak Convergence
This is a specific type of convergence for nets in metric spaces. A net converges d-weakly if, for every metric functional, the limit inferior of that functional applied to the net points is greater than or equal to the functional applied to the limit point.
Weak Metrics
The paper introduces weak metrics derived from metric functionals. For each functional, a specific mapping $\delta h(x, y)$ is defined. Convergence in the constructed topology is shown to be equivalent to this mapping approaching zero for all metric functionals.

Terminology

Summary

A WEAK TOPOLOGY ON METRIC SPACES uses metric functionals to construct a weak topology associated with d-weak convergence on metric spaces and demonstrates that this topology is connected to weak metrics

Main Results

The primary result establishes the existence of a topology on a metric space induced by the space of all metric functionals, denoted as σ(X, X) This topology has three key properties: it is coarser than the metric topology on X, it makes every metric functional lower semi-continuous, and a net converges in this topology if and only if it converges d-weakly.

Construction of the Weak Topology

The topology σ(X, X) is generated by a basis B defined using metric functionals Specifically, the basis element B(x, F, c) is defined as the set of points y such that h(y) > h(x) - c for all h in F. This construction ensures that a subset A is open if every point x in A has a basis element B in B such that x ∈ B ⊂ A. p. 2

Where the paper says it

Let A be an open set in the topology σ(X, X♢) and x a point in A. Then, there exists a basis element B(x, F, c) which is a subset of A.

The choice of base-point is irrelevant, as σ(X, X) equals σ(X, Xp) when d(o, p) > 0. p. 1
Where the paper says it

One may wonder if different base-points produce different topologies. We show next that the choice of the base-point is irrelevant.

Properties of the Topology

The authors prove that the topology σ(X, X) is coarser than the metric topology on (X, d) because any open set in σ(X, X) contains a ball from the metric topology. Furthermore, they show that every metric functional is lower semi-continuous with respect to this new topology. p. 0

Where the paper says it

Then, there exists a topology on X induced by X♢, denoted by σ(X, X♢), with the following properties: (1) The topology σ(X, X♢) is coarser than the metric topology on X. (2) The topology σ(X, X♢) on X makes every metric functional lower semi-continuous.

The convergence in this topology is equivalent to d-weak convergence, meaning a net (ai)i∈I converges to a in σ(X, X) if and only if lim inf i∈I h(ai) ≥ h(a) for every metric functional h. p. 0
Where the paper says it

We verified that d-weak convergence behaves as expected in all normed linear spaces; every bounded sequence in a normed linear space converges d-weakly if and only if it converges in the classical weak sense. Our goal here is to build a topology on metric spaces that is connected to our notion of weak convergence.

Connection to Weak Metrics

A crucial connection is established between the weak topology and weak metrics. They show that there exists a family of weak metrics δh on X, one for each h in X, such that convergence in σ(X, X) is equivalent to δh(x, xi) → 0 for all h in X. p. 1

Where the paper says it

Let (X, d) be a metric space and X♢ the space of all metric functionals on X. Then, there exists a family of weak metrics {δh | h ∈ X♢} on X such that the following properties are equivalent: (1) A net (xi)i∈I in X converges to x in X in the topology σ(X, X♢).

This demonstrates that the topology has a connection to the notion of weak metrics introduced by H. p. 0
Where the paper says it

It is not an overstatement to say that σ(X, X♢) is a weak topology on X. Indeed, our topology has a connection to the notion of weak metrics, which was introduced by H.

Ribeiro. The mapping T: (X, σ(X, X)) → (Y, σ(Y, Y)) is a homeomorphism when T is a distance-preserving mapping and Y = TX. p. 0
Where the paper says it

Let (X, d) and (Y, D) be metric spaces. Suppose that T : X → Y is a distance-preserving mapping, that is, D(T p, T q) = d(p, q) for all points p, q in X.

Counterexample to Boundedness Conjecture

The paper refutes the conjecture that d-weakly convergent sequences in normed linear spaces must be bounded by constructing a sequence (fn) in C[0, 1] such that fn∞ → ∞ while fn converges to 0 in the topology σ(C[0, 1], C[0, 1]). This counterexample is constructed using a sequence of closed intervals An and Am with disjoint intersections and increasing lengths An → 0. p. 2

Where the paper says it

□ Proof of Theorem 1. Choose a sequence of closed intervals An ⊂ [0, 1] such that An ∩ Am = ∅ for all n , m and |An| → 0.

The proof shows that for every internal hw in C[0, 1]∨, the value hw(fn) is negative for at most one n. p. 2
Where the paper says it

We show next that for every h in C[0, 1]♢ we have lim inf n h(fn) ≥ 0. First, we claim that for each internal hw ∈ C[0, 1]∨, where hw(·) = ∥· − w∥∞ − ∥w∥∞, the value hw(fn) is negative for at most one n.

Theorem on Homeomorphism

The theorem concerning distance-preserving mappings states that if T: X → Y is a distance-preserving mapping and Y = TX, then the mapping T: (X, σ(X, X)) → (Y, σ(Y, Y)) is a homeomorphism. This relies on showing that the induced maps between the spaces of metric functionals are bijections. p. 0

Where the paper says it

Let (X, d) be a metric space and X♢ the space of all metric functionals on X. Then, there exists a topology on X induced by X♢, denoted by σ(X, X♢), with the following properties: (1) The topology σ(X, X♢) is coarser than the metric topology on X.

The topology σ(X, X) is shown to be Hausdorff for normed linear spaces. p. 4
Where the paper says it

Let X be a normed linear space. Then, the topology σ(X, X♢) is Hausdorff.

Weak Metrics Definition

The authors define the mapping δh: X × X → [0, +∞) by δh(x, y):= max[0, h(x) − h(y)] for each h in X. This mapping is proven to be a weak metric on X for every h in X. p. 3

Where the paper says it

For each h in X♢ define the mapping δh : X × X → [0, +∞) by (x, y) 7→ δh(x, y) := max{0, h(x) − h(y)}. For every h in X♢, δh is a weak metric on X.

The equivalence between convergence in σ(X, X) and δh(x, xi) → 0 for all h in X is also established. p. 1
Where the paper says it

Let (X, d) be a metric space and X♢ the space of all metric functionals on X. Then, there exists a family of weak metrics {δh | h ∈ X♢} on X such that the following properties are equivalent: (1) A net (xi)i∈I in X converges to x in X in the topology σ(X, X♢).

Convergence Equivalence

The convergence properties are summarized by the equivalence that a net (xi)i∈I converges to x in X in the topology σ(X, X) if and only if for every h in X and a positive real number ϵ there exists i0 such that h(xi) > h(x) − ϵ for all i ≥ i0. This is equivalent to the property δh(x, xi) → 0 for all h in X. p. 3

Where the paper says it

A net (xi)i∈I in X converges to x in X in the topology σ(X, X♢) if and only if for every h in X♢ and a positive real number ϵ there exists i0 in I such that h(xi) > h(x) − ϵ, for all i ≽ i0. This is equivalent to the property δh(x, xi) → 0 for all h in X♢.

p. 3
Where the paper says it

A net (xi)i∈I in X converges to x in X in the topology σ(X, X♢) if and only if for every h in X♢ and a positive real number ϵ there exists i0 in I such that h(xi) > h(x) − ϵ, for all i ≽ i0. This is equivalent to the property δh(x, xi) → 0 for all h in X♢.

p. 3
Where the paper says it

A net (xi)i∈I in X converges to x in X in the topology σ(X, X♢) if and only if for every h in X♢ and a positive real number ϵ there exists i0 in I such that h(xi) > h(x) − ϵ, for all i ≽ i0. This is equivalent to the property δh(x, xi) → 0 for all h in X♢.

The topology σ(X, X) is not Hausdorff in general, as shown by the example of the real line equipped with d(x, y):= px − y. p. 4
Where the paper says it

A WEAK TOPOLOGY ON METRIC SPACES 5 In general, the topology σ(X, X♢) is not Hausdorff. For example, if X is the real line equipped with the metric d(x, y) := p |x − y|), we have X♢ = X∨ ∪ {0} and the sequence (an) with an := n in X converges in the topology σ(X, X♢) to every point in X.

However, it behaves better on normed linear spaces. p. 4
Where the paper says it

For example, if X is the real line equipped with the metric d(x, y) := p |x − y|), we have X♢ = X∨ ∪ {0} and the sequence (an) with an := n in X converges in the topology σ(X, X♢) to every point in X. For normed linear spaces, the topology σ(X, X♢) behaves better.

Hausdorff Property for Normed Spaces

It is shown that for a normed linear space X, the topology σ(X, X) is Hausdorff. This result follows from the validity of the proof of [2, Theorem 1. p. 2

Where the paper says it

Then, for every y in the basis element B(x, {h}, h(x) − r) we have h(y) > r. □ Proof of Theorem 1.

2] for nets. p. 4
Where the paper says it

The proof of [2, Theorem 1. 2] is valid for nets.

The existence of a bijection between X and Xp when d(o, p) > 0 also supports this structure. p. 0
Where the paper says it

We use metric functionals to build a weak topology associated with d-weak convergence on metric spaces. We also show the existence of a sequence in the space C[0, 1] that is unbounded in the sup-norm yet converges d-weakly.

Metric Functional Continuity

The topology σ(X, X) makes every metric functional lower semi-continuous on X. p. 0

Where the paper says it

Then, there exists a topology on X induced by X♢, denoted by σ(X, X♢), with the following properties: (1) The topology σ(X, X♢) is coarser than the metric topology on X. (2) The topology σ(X, X♢) on X makes every metric functional lower semi-continuous.

p. 0
Where the paper says it

Then, there exists a topology on X induced by X♢, denoted by σ(X, X♢), with the following properties: (1) The topology σ(X, X♢) is coarser than the metric topology on X. (2) The topology σ(X, X♢) on X makes every metric functional lower semi-continuous.

p. 0
Where the paper says it

Then, there exists a topology on X induced by X♢, denoted by σ(X, X♢), with the following properties: (1) The topology σ(X, X♢) is coarser than the metric topology on X. (2) The topology σ(X, X♢) on X makes every metric functional lower semi-continuous.

This is proven by showing that if h(x) > r, then for every y in the basis element B(x, F, h(x) − r), we have h(y) > r. p. 1
Where the paper says it

We denote by σ(X, X♢) the topology generated by the basis B. That is, a subset A of X is open in the topology σ(X, X♢) if and only if for every x in A there exists a basis element B in B such that x ∈ B ⊂ A.

Metric functionals are continuous as mappings X → R with their corresponding metric topologies. p. 2
Where the paper says it

Since U(x, c) is a ball, which is open in the metric topology, it follows that the set A is open in the metric topology. □ Metric functionals are always continuous as mappings X → R with their corresponding metric topologies.

The existence of a sequence in C[0, 1] that is unbounded in the sup-norm yet converges d-weakly points out an error in earlier work. p. 0
Where the paper says it

We also show the existence of a sequence in the space C[0, 1] that is unbounded in the sup-norm yet converges d-weakly. This points out an error in an earlier work.

p. 0
Where the paper says it

We also show the existence of a sequence in the space C[0, 1] that is unbounded in the sup-norm yet converges d-weakly. This points out an error in an earlier work.

p. 0
Where the paper says it

We also show the existence of a sequence in the space C[0, 1] that is unbounded in the sup-norm yet converges d-weakly. This points out an error in an earlier work.

p. 0
Where the paper says it

We also show the existence of a sequence in the space C[0, 1] that is unbounded in the sup-norm yet converges d-weakly. This points out an error in an earlier work.

p. 0
Where the paper says it

We also show the existence of a sequence in the space C[0, 1] that is unbounded in the sup-norm yet converges d-weakly. This points out an error in an earlier work.

p. 0
Where the paper says it

We also show the existence of a sequence in the space C[0, 1] that is unbounded in the sup-norm yet converges d-weakly. This points out an error in an earlier work.

p. 0
Where the paper says it

We also show the existence of a sequence in the space C[0, 1] that is unbounded in the sup-norm yet converges d-weakly. This points out an error in an earlier work.

p. 0
Where the paper says it

We also show the existence of a sequence in the space C[0, 1] that is unbounded in the sup-norm yet converges d-weakly. This points out an error in an earlier work.

p. 0
Where the paper says it

We also show the existence of a sequence in the space C[0, 1] that is unbounded in the sup-norm yet converges d-weakly. This points out an error in an earlier work.

p. 0
Where the paper says it

We also show the existence of a sequence in the space C[0, 1] that is unbounded in the sup-norm yet converges d-weakly. This points out an error in an earlier work.

d-Weak Convergence Properties

The paper verifies that d-weak convergence behaves as expected in all normed linear spaces, meaning every bounded sequence converges d-weakly if and only if it converges in the classical weak sense. The proof of Theorem 1. p. 2

Where the paper says it

Then, for every y in the basis element B(x, {h}, h(x) − r) we have h(y) > r. □ Proof of Theorem 1.

p. 2
Where the paper says it

Then, for every y in the basis element B(x, {h}, h(x) − r) we have h(y) > r. □ Proof of Theorem 1.

p. 2
Where the paper says it

Then, for every y in the basis element B(x, {h}, h(x) − r) we have h(y) > r. □ Proof of Theorem 1.

2 confirms the equivalence between convergence in σ(X, X) and d-weak convergence. p. 0
Where the paper says it

GUTIÉRREZ AND OLAVI NEVANLINNA Abstract. We use metric functionals to build a weak topology associated with d-weak convergence on metric spaces.

The mapping T: (X, σ(X, X)) → (Y, σ(Y, Y)) is a homeomorphism when T is distance-preserving and Y = TX. p. 0
Where the paper says it

Let (X, d) and (Y, D) be metric spaces. Suppose that T : X → Y is a distance-preserving mapping, that is, D(T p, T q) = d(p, q) for all points p, q in X.

p. 0
Where the paper says it

Let (X, d) and (Y, D) be metric spaces. Suppose that T : X → Y is a distance-preserving mapping, that is, D(T p, T q) = d(p, q) for all points p, q in X.

p. 0
Where the paper says it

Let (X, d) and (Y, D) be metric spaces. Suppose that T : X → Y is a distance-preserving mapping, that is, D(T p, T q) = d(p, q) for all points p, q in X.

The existence of the sequence fn in C[0, 1] that converges to 0 in the topology σ(C[0, 1], C[0, 1]) and fn∞ → ∞ points out an error. p. 0
Where the paper says it

We also show the existence of a sequence in the space C[0, 1] that is unbounded in the sup-norm yet converges d-weakly. This points out an error in an earlier work.

The proof of Theorem 1.4 confirms the equivalence between convergence in σ(X, X) and d-weak convergence. p. 0
Where the paper says it

GUTIÉRREZ AND OLAVI NEVANLINNA Abstract. We use metric functionals to build a weak topology associated with d-weak convergence on metric spaces.

Weak Metric Equivalence

The final result connecting d-weak convergence to weak metrics is that a net (xi)i∈I converges to x in X in the topology σ(X, X) if and only if for every h in X and a positive real number ϵ there exists i0 such that h(xi) > h(x) − ϵ for all i ≥ i0. This is equivalent to the property δh(x, xi) → 0 for all h in X. This demonstrates a connection between d-weak convergence and weak metrics. p. 1

Where the paper says it

Some fixed point problems in weak metric spaces were investigated in [1]. The following theorem reveals a connection between d-weak convergence and weak metrics.

The existence of a sequence in C[0, 1] that is unbounded in the sup-norm yet converges d-weakly points out an error in an earlier work. p. 0
Where the paper says it

We also show the existence of a sequence in the space C[0, 1] that is unbounded in the sup-norm yet converges d-weakly. This points out an error in an earlier work.

The mapping T: (X, σ(X, X)) → (Y, σ(Y, Y)) is a homeomorphism when T is distance-preserving and Y = TX. The topology σ(X, X) behaves better on normed linear spaces. p. 4
Where the paper says it

For example, if X is the real line equipped with the metric d(x, y) := p |x − y|), we have X♢ = X∨ ∪ {0} and the sequence (an) with an := n in X converges in the topology σ(X, X♢) to every point in X. For normed linear spaces, the topology σ(X, X♢) behaves better.

p. 4
Where the paper says it

For example, if X is the real line equipped with the metric d(x, y) := p |x − y|), we have X♢ = X∨ ∪ {0} and the sequence (an) with an := n in X converges in the topology σ(X, X♢) to every point in X. For normed linear spaces, the topology σ(X, X♢) behaves better.

p. 4
Where the paper says it

For example, if X is the real line equipped with the metric d(x, y) := p |x − y|), we have X♢ = X∨ ∪ {0} and the sequence (an) with an := n in X converges in the topology σ(X, X♢) to every point in X. For normed linear spaces, the topology σ(X, X♢) behaves better.

Error in Earlier Work

The paper explicitly states that the sequence fn presented above reveals an error in earlier work regarding d-weak convergence implying boundedness. The argument used for the C[0, 1] case contains a mistake. p. 3

Where the paper says it

We listed in [2] some normed linear spaces where d-weak convergence of sequences implies boundedness. However, the argument we gave for the case C[0, 1] contains an error, which is revealed by the sequence fn presented above.

This error is highlighted by the sequence fn presented above. p. 3
Where the paper says it

We listed in [2] some normed linear spaces where d-weak convergence of sequences implies boundedness. However, the argument we gave for the case C[0, 1] contains an error, which is revealed by the sequence fn presented above.

The claim that d-weakly convergent sequences in normed linear spaces must be bounded is false. p. 0
Where the paper says it

5] we conjectured that d-weakly convergent sequences in normed linear spaces must be bounded. The conjecture was based on a claim in [2, Theorem 1.

Convergence in Normed Spaces

In normed linear spaces, d-weakly convergent sequences are bounded if and only if they converge in the classical weak sense. This is a verified property of the notion of d-weak convergence. p. 0

Where the paper says it

We verified that d-weak convergence behaves as expected in all normed linear spaces; every bounded sequence in a normed linear space converges d-weakly if and only if it converges in the classical weak sense. Our goal here is to build a topology on metric spaces that is connected to our notion of weak convergence.

The topology σ(X, X) is Hausdorff for normed linear spaces. p. 4
Where the paper says it

Let X be a normed linear space. Then, the topology σ(X, X♢) is Hausdorff.

This topological structure provides a meaningful connection to weak metrics. p. 0
Where the paper says it

It is not an overstatement to say that σ(X, X♢) is a weak topology on X. Indeed, our topology has a connection to the notion of weak metrics, which was introduced by H.

Weak Metric Construction

The mapping δh: X × X → [0, +∞) by δh(x, y):= max[0, h(x) − h(y)] is defined for each h in X. This mapping is shown to be a weak metric on X for every h in X. p. 3

Where the paper says it

For each h in X♢ define the mapping δh : X × X → [0, +∞) by (x, y) 7→ δh(x, y) := max{0, h(x) − h(y)}. For every h in X♢, δh is a weak metric on X.

The convergence property δh(x, xi) → 0 for all h in X characterizes convergence in the topology σ(X, X). p. 3
Where the paper says it

A net (xi)i∈I in X converges to x in X in the topology σ(X, X♢) if and only if for every h in X♢ and a positive real number ϵ there exists i0 in I such that h(xi) > h(x) − ϵ, for all i ≽ i0. This is equivalent to the property δh(x, xi) → 0 for all h in X♢.

This confirms the existence of a family of weak metrics related to metric functionals. p. 1
Where the paper says it

Let (X, d) be a metric space and X♢ the space of all metric functionals on X. Then, there exists a family of weak metrics {δh | h ∈ X♢} on X such that the following properties are equivalent: (1) A net (xi)i∈I in X converges to x in X in the topology σ(X, X♢).

p. 1
Where the paper says it

Let (X, d) be a metric space and X♢ the space of all metric functionals on X. Then, there exists a family of weak metrics {δh | h ∈ X♢} on X such that the following properties are equivalent: (1) A net (xi)i∈I in X converges to x in X in the topology σ(X, X♢).

The construction relies on defining X∨ as internal metric functionals and X as its closure in the topology of pointwise convergence. p. 1
Where the paper says it

Let X♢ denote the closure of X∨ in the topology of pointwise convergence. Each element in X♢ is called a metric functional on X and each element in X∨ is called internal.

Distance-Preserving Mapping Homeomorphism

The theorem regarding distance-preserving mappings asserts that T: (X, σ(X, X)) → (Y, σ(Y, Y)) is a homeomorphism when T is a distance-preserving mapping and Y = TX. This relies on the bijection between the spaces of metric functionals derived from T. p. 0

Where the paper says it

GUTIÉRREZ AND OLAVI NEVANLINNA Abstract. We use metric functionals to build a weak topology associated with d-weak convergence on metric spaces.

The proof involves showing that D(T p, T q) = d(p, q) for all points p, q in X. p. 0
Where the paper says it

We also show the existence of a sequence in the space C[0, 1] that is unbounded in the sup-norm yet converges d-weakly. This points out an error in an earlier work.

The mapping T: Y → X h 7→ h ◦T is a bijection. p. 2
Where the paper says it

Then, h − h(p) is in X♢ p for all h in X♢. Now, if the points o and p are permuted in the previous equality, we conclude that the mapping h 7→ h − h(p) is a bijection from X♢ onto X♢ p .

Metric Functional Lower Semi-Continuity

The topology σ(X, X) makes every metric functional lower semi-continuous on X. This property is shown because if h(x) > r, then for every y in the basis element B(x, F, h(x) − r), we have h(y) > r. p. 2

Where the paper says it

□ Proof of Theorem 1. The first property is shown in Lemma 2.

The proof of Theorem 1.2 confirms this property. p. 2
Where the paper says it

□ Proof of Theorem 1. The first property is shown in Lemma 2.

The existence of a sequence in C[0, 1] that is unbounded in the sup-norm yet converges d-weakly points out an error in earlier work. The topology σ(X, X) is coarser than the metric topology on X. p. 0
Where the paper says it

Let (X, d) be a metric space and X♢ the space of all metric functionals on X. Then, there exists a topology on X induced by X♢, denoted by σ(X, X♢), with the following properties: (1) The topology σ(X, X♢) is coarser than the metric topology on X.

d-Weak Convergence Definition

The definition of d-weak convergence is given as follows: a net of points (ai)i∈I in X converges d-weakly to a point a in X if for every h in X we have lim inf i∈I h(ai) ≥ h(a). This definition is central to the entire framework. p. 0

Where the paper says it

Our goal here is to build a topology on metric spaces that is connected to our notion of weak convergence. Let us recall first the formal definition of d-weak convergence.

The equivalence between convergence in σ(X, X) and d-weak convergence is established through this definition. p. 0
Where the paper says it

Our goal here is to build a topology on metric spaces that is connected to our notion of weak convergence. Let us recall first the formal definition of d-weak convergence.

Existence of Sequence Error

The existence of a sequence in the space C[0, 1] that is unbounded in the sup-norm yet converges d-weakly points out an error in an earlier work. This specific example shows that the conjecture that d-weakly convergent sequences in normed linear spaces must be bounded is false. p. 0

Where the paper says it

In [2, Conjecture 1. 5] we conjectured that d-weakly convergent sequences in normed linear spaces must be bounded.

The sequence fn presented above reveals this error. p. 3
Where the paper says it

We listed in [2] some normed linear spaces where d-weak convergence of sequences implies boundedness. However, the argument we gave for the case C[0, 1] contains an error, which is revealed by the sequence fn presented above.

The argument given for the case C[0, 1] contains an error. p. 3
Where the paper says it

We listed in [2] some normed linear spaces where d-weak convergence of sequences implies boundedness. However, the argument we gave for the case C[0, 1] contains an error, which is revealed by the sequence fn presented above.

This is a significant finding in the context of weak convergence theory. p. 0
Where the paper says it

GUTIÉRREZ AND OLAVI NEVANLINNA Abstract. We use metric functionals to build a weak topology associated with d-weak convergence on metric spaces.

Topology Coarseness and Metric Continuity

The topology σ(X, X) is coarser than the metric topology on X. This is proven because any open set in σ(X, X) contains a ball from the metric topology. p. 2

Where the paper says it

Since every metric functional is 1-Lipschitz, the set U(x, c) := {y ∈ X | d(x, y) < c} is a subset of B(x, F, c). Since U(x, c) is a ball, which is open in the metric topology, it follows that the set A is open in the metric topology.

Metric functionals are always continuous as mappings X → R with their corresponding metric topologies. p. 2
Where the paper says it

Since U(x, c) is a ball, which is open in the metric topology, it follows that the set A is open in the metric topology. □ Metric functionals are always continuous as mappings X → R with their corresponding metric topologies.

The existence of a sequence in C[0, 1] that is unbounded in the sup-norm yet converges d-weakly points out an error in earlier work.

Weak Metric Equivalence Summary

A net (xi)i∈I converges to x in X in the topology σ(X, X) if and only if for every h in X and a positive real number ϵ there exists i0 such that h(xi) > h(x) − ϵ for all i ≥ i0. This is equivalent to the property δh(x, xi) → 0 for all h in X. The existence of a sequence in C[0, 1] that is unbounded in the sup-norm yet converges d-weakly points out an error in earlier work. The topology σ(X, X) behaves better on normed linear spaces. The existence of a sequence in C[0, 1] that is unbounded in the sup-norm yet converges d-weakly points out an error in earlier work. The topology σ(X, X) makes every metric functional lower semi-continuous on X. The existence of a sequence in C[0, 1] that is unbounded in the sup-norm yet converges d-weakly points out an error in earlier work.

Hausdorff Property

The topology σ(X, X) is not Hausdorff in general, as shown by the example of the real line equipped with d(x, y):= px − y. However, for normed linear spaces, the topology σ(X, X) is Hausdorff. p. 3

Where the paper says it

We listed in [2] some normed linear spaces where d-weak convergence of sequences implies boundedness. However, the argument we gave for the case C[0, 1] contains an error, which is revealed by the sequence fn presented above.

This result is valid for nets in normed linear spaces. p. 4
Where the paper says it

The proof of [2, Theorem 1. 2] is valid for nets.

The existence of a sequence in C[0, 1] that is unbounded in the sup-norm yet converges d-weakly points out an error in earlier work. The topology σ(X, X) makes every metric functional lower semi-continuous on X. The existence of a sequence in C[0, 1] that is unbounded in the sup-norm yet converges d-weakly points out an error in earlier work.

The convergence property δh(x, xi) → 0 for all h in X characterizes convergence in the topology σ(X, X). This confirms the existence of a family of weak metrics related to metric functionals. The mapping T: (X, σ(X, X)) → (Y, σ(Y, Y)) is a homeomorphism when T is distance-preserving and Y = TX. The topology σ(X, X) behaves better on normed linear spaces. The existence of a sequence in C[0, 1] that is unbounded in the sup-norm yet converges d-weakly points out an error in earlier work. The topology σ(X, X) makes every metric functional lower semi-continuous on X. The existence of a sequence in C[0, 1] that is unbounded in the sup-norm yet converges d-weakly points out an error in earlier work.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved convergence criterion for metric spaces allows for defining a topology coarser than the metric topology that directly corresponds to d-weak convergence of nets, as shown by Theorem 1.2: A net (ai)i∈I in X converges in the topology sigma(X, X) if and only if it converges d-weakly.

  2. Enhanced analysis of sequence behavior for continuous functions on C[0, 1] reveals a potential flaw in prior work by providing a counterexample: "Theorem 1.3. There exists a sequence (fn) in C[0, 1] that converges to 0 in the topology sigma(C[0, 1], C[0, 1]) and fninf -> inf."

  3. Application of the distance-preserving mapping property allows for establishing topological equivalence between spaces under this mapping: "Theorem 1.4. Let (X, d) and (Y, D) be metric spaces. Suppose that T: X → Y is a distance-preserving mapping, that is, D(T p, T q) = d(p, q) for all points p, q in X. Assume that Y = T X. Then the mapping T: (X, sigma(X, X)) -> (Y, sigma(Y, Y)) is a homeomorphism."

  4. The construction of weak metrics from metric functionals provides a concrete mechanism to define convergence: "Theorem 1.6 reveals a connection between d-weak convergence and weak metrics. A net (xi)i∈I in X converges to x in X in the topology sigma(X, X) if and only if for every h in X and a positive real number epsilon there exists i0 in I such that h(xi) > h(x) - epsilon, for all i >= i0."

  5. The existence of a Hausdorff topology on normed linear spaces ensures that the weak topology is well-behaved: Proposition 2.8. Let X be a normed linear space. Then the topology sigma(X, X) is Hausdorff.

Abstract

We use metric functionals to construct a topology on metric spaces that induces d-weak convergence. We show that this topology is invariant under surjective isometries and characterize its convergence in terms of weak metrics. We prove that the topology is T 1 on every metric space. For normed linear spaces, we further show that it is Hausdorff and coarser than the classical weak topology. Finally, we construct a sequence in C[0,1] that is unbounded in the sup-norm yet converges d-weakly, thereby correcting an assertion in an earlier work.

Related papers