Chapter 108. Functional Analysis Connections

Functional analysis studies vector spaces together with topology, norms, inner products, and linear operators. It extends linear algebra from finite-dimensional coordinate spaces to spaces of functions, sequences, distributions, and operators.

The central objects are Banach spaces, Hilbert spaces, and continuous linear operators. A Banach space is a complete normed vector space. A Hilbert space is a complete inner product space. Functional analysis studies these spaces and the bounded linear maps between them.

108.1 From Linear Algebra to Functional Analysis

Finite-dimensional linear algebra studies spaces such as

Functional analysis studies spaces such as

The objects are still vectors. A function can be added to another function. A sequence can be multiplied by a scalar. A differential equation can be viewed as an equation in a vector space.

The new issue is convergence.

In finite dimensions, algebra and topology are tightly linked. In infinite dimensions, they separate. A linear map may fail to be continuous. A subspace may fail to be closed. A bounded sequence may fail to have a convergent subsequence.

Functional analysis studies linear algebra under these analytic constraints.

108.2 Normed Spaces

A normed vector space is a vector space $V$ with a function

satisfying:

and

The norm gives a notion of length. It also defines distance:

Thus a normed space is both algebraic and metric.

Examples include:

108.3 Banach Spaces

A Banach space is a complete normed vector space.

Completeness means every Cauchy sequence converges to an element of the space.

If

is Cauchy, then for every $\varepsilon>0$, there exists $N$ such that

for all $m,n\ge N$.

The space is complete if there exists $v\in V$ such that

Completeness is essential because analysis repeatedly constructs objects as limits. A space used for solving equations should contain the limits produced by its approximation methods.

Standard Banach spaces include $C([a,b])$ with the supremum norm, $\ell^p$ for $1\le p\le\infty$, and $L^p$ spaces for $1\le p\le\infty$. Functional analysis treats continuous linear operators on Banach and Hilbert spaces as central objects.

108.4 Inner Product Spaces

An inner product space is a vector space with a pairing

that generalizes the dot product.

Over $\mathbb R$, it satisfies:

and

for $v\ne 0$.

The inner product defines a norm:

It also defines orthogonality:

Inner product spaces generalize Euclidean geometry and provide formal notions of length, angle, and projection.

108.5 Hilbert Spaces

A Hilbert space is a complete inner product space.

Every Hilbert space is a Banach space because its inner product defines a norm. Hilbert spaces behave more like Euclidean spaces than general Banach spaces because they support orthogonality and projection.

Examples include:

Hilbert spaces are the natural setting for Fourier analysis, quantum mechanics, spectral theory, and weak formulations of differential equations.

108.6 Bounded Linear Operators

Let $X$ and $Y$ be normed spaces.

A linear operator

is bounded if there exists $C\ge 0$ such that

for all $x\in X$.

The least such constant is the operator norm:

For linear maps between normed spaces, boundedness is equivalent to continuity.

This is one of the basic bridges between algebra and analysis. A bounded linear operator is a linear map compatible with the topology.

108.7 Operator Spaces

The set of bounded linear operators from $X$ to $Y$ is denoted

It is itself a normed vector space under the operator norm.

When $Y$ is complete, $\mathcal B(X,Y)$ is also complete. Thus if $Y$ is a Banach space, then $\mathcal B(X,Y)$ is a Banach space.

When $X=Y$, one writes

This space is an algebra under composition:

Thus functional analysis studies not only vector spaces, but also algebras of operators on vector spaces.

108.8 Linear Functionals

A linear functional is a linear map

If $X$ is normed, one usually studies continuous linear functionals.

The continuous dual space is

It consists of all bounded linear functionals on $X$.

The dual space is one of the main tools of functional analysis. It allows vectors to be studied through scalar measurements.

For finite-dimensional spaces, the algebraic dual and continuous dual coincide. In infinite-dimensional normed spaces, they may differ greatly.

108.9 Hahn-Banach Theorem

The Hahn-Banach theorem is one of the foundational theorems of functional analysis.

In one common form, it says that a bounded linear functional defined on a subspace can be extended to the whole normed space without increasing its norm.

Thus, if

and

is bounded and linear, then under the usual hypotheses there exists a bounded linear extension

such that

and

The theorem guarantees that normed spaces have enough continuous linear functionals to separate points and study dual spaces.

108.10 Duality

Duality is the study of a space through its linear functionals.

A vector $x\in X$ can be tested by functionals:

If sufficiently many functionals are available, the behavior of $x$ can be recovered from all such scalar evaluations.

For example, in a Hilbert space $H$, the Riesz representation theorem states that every continuous linear functional has the form

for a unique $y\in H$.

Thus

in a canonical way, up to the convention for complex conjugation.

This is one reason Hilbert spaces have especially strong geometric structure.

108.11 Weak Convergence

Norm convergence means

Weak convergence means that all continuous linear functionals converge on the sequence:

for every

Weak convergence is written

Weak convergence is weaker than norm convergence. It records convergence under all scalar tests, but not necessarily convergence in length.

Weak topologies are important because they often provide compactness where norm topology does not.

108.12 Banach-Alaoglu Theorem

The Banach-Alaoglu theorem states that the closed unit ball of the dual space of a normed vector space is compact in the weak star topology.

This theorem is a substitute for the finite-dimensional fact that closed bounded sets are compact.

In infinite-dimensional normed spaces, the closed unit ball usually fails to be compact in the norm topology. Weak and weak star topologies recover compactness in a form useful for analysis.

This compactness principle is central in optimization, PDEs, measure theory, and variational methods.

108.13 Open Mapping Theorem

The open mapping theorem states that a surjective bounded linear operator between Banach spaces maps open sets to open sets. One important consequence is the bounded inverse theorem: a bijective bounded linear operator between Banach spaces has a bounded inverse.

This result has no purely algebraic analogue.

Algebra says that a bijective linear map has an inverse. Functional analysis asks whether the inverse is continuous.

For Banach spaces, completeness makes the answer positive.

108.14 Closed Graph Theorem

The closed graph theorem gives another criterion for continuity.

If

is a linear operator between Banach spaces, then $T$ is continuous if and only if its graph

is closed in

This theorem connects an analytic property of an operator with a topological property of its graph.

It is especially useful when continuity is hard to prove directly.

108.15 Uniform Boundedness Principle

The uniform boundedness principle is another central theorem of Banach-space theory.

It says that a pointwise bounded family of bounded linear operators on a Banach space is uniformly bounded in operator norm.

That is, if

for every fixed $x\in X$, then

The theorem turns pointwise control into uniform operator control.

It is one of the main tools for detecting hidden unboundedness in approximation processes.

108.16 Orthogonal Projection

Hilbert spaces support orthogonal projection.

If $M$ is a closed subspace of a Hilbert space $H$, then every $x\in H$ can be written uniquely as

where

The vector $m$ is the orthogonal projection of $x$ onto $M$.

This generalizes the projection theorem from Euclidean geometry.

The requirement that $M$ be closed is essential. In infinite dimensions, nonclosed subspaces may fail to contain best approximations.

108.17 Spectral Theory

Spectral theory generalizes eigenvalue theory.

For a linear operator $T$, a scalar $\lambda$ belongs to the spectrum if

fails to have a bounded inverse.

In finite dimensions, this is equivalent to

In infinite dimensions, determinants are usually unavailable, and the spectrum may contain values that are not eigenvalues.

Thus functional analysis replaces the characteristic polynomial with operator-theoretic invertibility.

Spectral theory is central in quantum mechanics, PDEs, numerical analysis, and signal processing.

108.18 Compact Operators

A compact operator sends bounded sets to relatively compact sets.

Compact operators are important because their spectral behavior resembles finite-dimensional linear algebra more closely than general bounded operators.

For many compact operators on infinite-dimensional Hilbert spaces, nonzero spectral values appear as eigenvalues with finite multiplicity and can accumulate only at zero.

Integral operators often provide examples:

Compactness turns an infinite-dimensional operator into something that can often be approximated effectively by finite-rank operators.

108.19 Self-Adjoint Operators

In a Hilbert space, the adjoint of an operator $T$ is an operator $T^*$ satisfying

An operator is self-adjoint if

Self-adjoint operators generalize real symmetric matrices and Hermitian matrices.

They have real spectral behavior and support spectral decompositions. They are central in quantum mechanics, where observables are modeled by self-adjoint operators.

108.20 Weak Formulations of Differential Equations

Functional analysis is a natural language for differential equations.

A differential equation can often be rewritten as an operator equation:

Instead of requiring classical derivatives, one may seek weak solutions in a Hilbert or Banach space.

For example, the equation

can be studied through the bilinear form

One then seeks $u$ such that

for all test functions $v$.

This converts a differential equation into a problem about linear functionals, bilinear forms, and Hilbert-space geometry.

108.21 Approximation and Projection

Functional analysis provides the foundation for approximation methods.

A common strategy is to choose finite-dimensional subspaces

and solve finite-dimensional problems in $V_n$.

This leads to Galerkin methods, finite element methods, spectral methods, and projection algorithms.

The linear algebra is finite at each stage. The functional analysis proves convergence as

Thus numerical analysis of infinite-dimensional problems depends on both matrix computation and analytic estimates.

108.22 Relation to Linear Algebra

Functional analysis can be viewed as linear algebra plus limits.

The finite-dimensional theory remains the model, but many statements require new hypotheses.

Completeness, continuity, boundedness, and compactness become structural assumptions rather than automatic facts.

108.23 Example: Shift Operator on $\ell^2$

Let

be defined by

This is the right shift operator.

It is linear and bounded.

Indeed,

Thus

The operator preserves norm, but it is not surjective. No vector maps to a sequence whose first coordinate is nonzero.

This example shows how simple infinite-dimensional operators can behave differently from square matrices.

108.24 Example: Multiplication Operator

Let

with the supremum norm, and define

Then $T$ is linear.

Also,

Thus

Taking $f(x)=1$, we get

so

This operator behaves like an infinite-dimensional diagonal matrix whose diagonal entries vary continuously over $[0,1]$.

108.25 Example: Integral Operator

Let

where $K$ is a continuous function on $[0,1]\times[0,1]$.

Such an operator maps functions to functions.

It is linear because integration is linear.

Under standard norms, integral operators are often bounded, and many are compact.

This makes them important in integral equations, PDEs, probability, and applied mathematics.

They are infinite-dimensional analogues of matrices, with the kernel $K(x,y)$ playing the role of matrix entries.

108.26 Why Functional Analysis Matters

Functional analysis matters because many problems are linear but infinite-dimensional.

Examples include:

The subject supplies the theorems that justify limiting processes, approximation schemes, and operator methods.

108.27 Summary

Functional analysis extends linear algebra to vector spaces with topology.

The main objects are:

The guiding principle is that linear algebra remains valid only when analytic structure is controlled. In finite dimensions, many properties are automatic. In infinite dimensions, they become theorems, hypotheses, or failures. Functional analysis is the discipline that studies this boundary.