When Are Span(X) Subsets Relatively Weakly Compact?

Hey guys, today we're diving deep into a super interesting question in the world of functional analysis and general topology: under what conditions is a subset of the Span($X$) relatively weakly compact? This is a pretty specific question, but it gets to the heart of understanding the structure of subsets within Banach spaces, especially when $X$ itself is a totally bounded subset. We're talking about really getting to grips with the nuances of weak compactness, which is a fundamental concept when you're dealing with function spaces, Hilbert spaces, and more abstract mathematical structures. So, grab your thinking caps, because we're going to unravel this one, piece by piece, making sure we cover all the juicy details.

Understanding the Core Concepts: Span($X$) and Weak Compactness

First off, let's break down what we mean by Span($X$). In linear algebra, the span of a set of vectors $X$ is the set of all possible finite linear combinations of those vectors. So, if $X = \{x_1, x_2, \dots, x_n\}$, then Span($X$) is the set of all vectors of the form $c_1 x_1 + c_2 x_2 + \dots + c_n x_n$, where $c_i$ are scalars. When $X$ is a subset of a Banach space, Span($X$) forms a linear subspace. Now, the problem specifies that $X$ is a totally bounded subset of a Banach space. This is a crucial piece of information. A set is totally bounded if, for any $\epsilon > 0$, it can be covered by a finite number of open balls of radius $\epsilon$. This is a stronger condition than just being bounded; it implies that the set is, in a sense, 'small' or 'dense' within its closure. Think of it as a sort of finite approximation property.

So, we're looking at the linear combinations of vectors from a 'small' set $X$. The question is about relative weak compactness. A subset $K$ of a topological vector space is relatively weakly compact if its closure in the weak topology, $\overline{K}^w$, is weakly compact. A set is weakly compact if it is a weakly closed and weakly bounded subset of a Banach space. For subsets of a Banach space, the Eberlein-\v{S}mulian theorem tells us that a subset is weakly compact if and only if it is weakly closed and sequentially weakly compact (meaning every sequence in the set has a weakly convergent subsequence). When we talk about relative weak compactness, we're interested in whether the closure of our set in the weak topology is itself weakly compact. This is super important because many powerful theorems in functional analysis, like the Dunford-Pettis theorem concerning the weak compactness of the image of a weakly compact set under an operator, rely on these notions.

Connecting Totally Bounded Sets to Weak Compactness

Now, let's connect these ideas. We have $X$ being totally bounded, and we're considering Span($X$). What can we say about the weak compactness of subsets within Span($X$)? A key result that often comes up in these discussions is that in a normed vector space, if a set is totally bounded, then its closure is compact in the norm topology. However, we're dealing with the weak topology, which is generally much coarser than the norm topology. The relationship between total boundedness and weak compactness is not always direct, but it's a fruitful area of investigation. For instance, if $X$ is a totally bounded subset of a Banach space $E$, then its closure $\overline{X}$ is a compact subset in the norm topology. If $E$ itself is a Hilbert space, then any closed and bounded subset is weakly compact. However, $E$ is just a general Banach space here.

Let's consider a subset $S \subseteq \text{Span}(X)$. For $S$ to be relatively weakly compact, its closure $\overline{S}^w$ must be weakly compact. A fundamental result for weak compactness in Banach spaces is the Pettis criterion: a subset $A$ of a Banach space $E$ is relatively weakly compact if and only if for every $\epsilon > 0$, there exists a weakly compact subset $K \subseteq E$ such that $A \subseteq K + \epsilon B_E$, where $B_E$ is the closed unit ball in $E$. Another crucial characterization relates to sequences: a subset $A$ is relatively weakly compact if and only if every sequence in $A$ has a weakly convergent subsequence.

The Role of the Banach Space Structure

Crucially, the structure of the Banach space $E$ plays a significant role. If $E$ is a Hilbert space, then the weak topology is closely related to the norm topology. In a Hilbert space, a subset is weakly compact if and only if it is closed and bounded. Since $X$ is totally bounded, its closure $\overline{X}$ is norm-compact. Span($X$) is the set of finite linear combinations of elements in $X$. If $X$ is finite, then Span($X$) is finite-dimensional, and thus any subset of Span($X$) is relatively norm-compact, which implies relative weak compactness. However, $X$ can be infinite.

If $E$ is reflexive, then the unit ball $B_E$ is weakly compact. The space $E^*$ is also reflexive. If $E$ is a dual space, the Banach-Alaoglu theorem states that the closed unit ball $B_E$ is weak-* compact. The weak topology on $E$ is given by the functionals in $E^*$. The weak topology on $E^*$ is the weak-* topology.

A Deeper Dive: The Pettis Integral and Bochner Spaces

Your specific problem mentions a Pettis-integrable Hilbert space-valued function $v: \Omega \to \mathcal H$, defined on a perfect probability space $(\Omega,

${}$\mathcal F, P)$. This context is important because the Pettis integral is a generalization of the Lebesgue integral for vector-valued functions, and it's intimately connected with weak compactness and convergence. A function $v$ is Pettis integrable if it is weakly measurable and its integral $\int_{\Omega} v dP$ exists in the weak sense, meaning $\phi \circ v$ is Lebesgue integrable for all $\phi$ in the dual space of the target space. The Pettis integral is crucial for establishing convergence theorems and understanding the behavior of functions in function spaces like $L^p$ spaces, especially Bochner spaces.

Consider the set $V = \{v(\omega) : \omega \in \Omega\}$. If $V$ is totally bounded in $\mathcal H$, then Span($V$) has certain properties. The set of all Pettis-integrable functions from $\Omega$ to a Banach space $E$ forms a vector space. If $E$ is a Hilbert space, it's a standard result that if $v: \Omega \to E$ is Pettis integrable and $v(\Omega)$ is totally bounded in $E$, then the image of the function $v$ is relatively weakly compact in $E$. That is, $v(\Omega)$ is relatively weakly compact. This implies that Span($v(\Omega)$), the set of finite linear combinations of the values of $v$, also has connections to weak compactness.

Specifically, if $v(\Omega)$ is totally bounded, then for any $\epsilon > 0$, there exists a finite set $\{y_1, \dots, y_n\} \subseteq v(\Omega)$ such that $v(\Omega) \subseteq \bigcup_{i=1}^n B(y_i, \epsilon)$. The set Span($v(\Omega)$) consists of all finite sums $\sum c_i v(\omega_i)$. If $v(\Omega)$ is relatively weakly compact, then its closure $\overline{v(\Omega)}^w$ is weakly compact. Since Span($v(\Omega)$) is a linear subspace, its closure $\overline{\text{Span}(v(\Omega))}^w$ is also weakly closed and thus weakly compact if $v(\\Omega)$ is itself contained in a weakly compact set.

Key Theorems and Conditions

Let $E$ be a Banach space and $X \subseteq E$ be a totally bounded subset. We are interested in the conditions under which a subset $S \subseteq \text{Span}(X)$ is relatively weakly compact.

1. If $E$ is a Hilbert space: In a Hilbert space $\mathcal H$, a subset is weakly compact if and only if it is closed and bounded. Since $X$ is totally bounded, its closure $\overline{X}$ is norm-compact. The closure of Span($X$) in the norm topology, $\overline{\text{Span}(X)}$, is the same as the closed span of $\overline{X}$. If $\overline{X}$ is norm-compact, then Span($\overline{X}$) is itself norm-compact (as it's the image of a compact set under a continuous map, or more accurately, the set of finite sums from a compact set is compact). Thus, $\overline{\text{Span}(X)}$ is norm-compact. Any norm-compact set is also weakly compact. So, if $X$ is totally bounded in a Hilbert space, then $\text{Span}(X)$ is contained in a weakly compact set, and therefore any subset $S \subseteq \text{Span}(X)$ is relatively weakly compact.

2. General Banach Spaces and the Role of Dual Spaces: For a general Banach space $E$, we need to be more careful. A subset $A \subseteq E$ is relatively weakly compact if and only if its closure in the weak topology, $\overline{A}^w$, is weakly compact. The Eberlein-\v{S}mulian theorem is key here: a subset of a Banach space is weakly compact iff it is sequentially weakly compact.

   A crucial condition for relative weak compactness is related to the dual space $E^*$. A subset $A o E$ is relatively weakly compact if and only if for every weakly convergent sequence $(x_n)$ in $E$, the sequence $(f(x_n))$ converges for every $f o E^*$. This is not quite right. The correct statement is that a subset $A$ of a Banach space $E$ is relatively weakly compact if and only if every sequence in $A$ has a subsequence that converges weakly to an element in $E$.

   A more direct condition involving the dual space is that a subset $A o E$ is relatively weakly compact if and only if its image under any weakly continuous functional $f o E^*$ is a relatively compact subset of $\mathbb{R}$. This is also not entirely accurate. The correct condition related to the dual space is that a subset $A$ is relatively weakly compact if and only if it is bounded and for every $\epsilon > 0$, there exists a finite-dimensional subspace $F \subseteq E$ such that $d(A, F) \leq \epsilon$, where $d(A, F)$ is the distance between $A$ and $F$ in the Hausdorff metric. This is related to the concept of the approximation property.

   Let's revisit the Pettis criterion. A set $A$ is relatively weakly compact if and only if it is bounded and for every $\epsilon > 0$, there exists a weakly compact set $K \subseteq E$ such that $A \subseteq K + \epsilon B_E$. If $X$ is totally bounded, then $\overline{X}$ is norm-compact. Since Span($X$) is the set of finite linear combinations of elements in $X$, and the set of finite linear combinations of elements in a norm-compact set is itself norm-compact, we have that $\overline{\text{Span}(X)}$ is norm-compact. Any norm-compact set is relatively weakly compact. Therefore, any subset $S$ of Span($X$) is relatively weakly compact if $X$ is a totally bounded subset of a Banach space $E$. This seems too general, so let's re-examine.

The Nuance of 'Subset' and 'Span'

Ah, the devil is in the details! The statement