Questions: Hahn-Banach Theorem

Question 1

The Hahn-Banach theorem guarantees that the dual space X* of a normed space X 'separates points.' What does this mean, and why is it significant?

• A. X* contains a basis for X, allowing every element to be expressed as a linear combination of functionals
• B. For any two distinct elements x₁ ≠ x₂ in X, there exists a bounded linear functional φ ∈ X* such that φ(x₁) ≠ φ(x₂); this guarantees the dual is rich enough to distinguish every element of X
• C. Every element of X can be identified with a unique functional in X* via a norm-preserving isomorphism
• D. The dual contains a single functional that attains its norm on every element of X simultaneously

Answer: (B)

Point separation means the dual space is rich enough to tell any two distinct elements apart. This is the practical content of 'the dual is large': given x₁ ≠ x₂, you can always find a bounded linear functional that evaluates differently on them. Without this, the dual might contain only the zero functional — useless for distinguishing elements — and the theories of weak convergence (which are defined using duals), reflexivity, and adjoint operators would collapse. Point separation is the minimum condition for doing analysis with dual spaces, and Hahn-Banach certifies it holds.

Question 2

Why does the Hahn-Banach extension proof require Zorn's lemma in infinite-dimensional spaces, when extending a linear functional is straightforward in finite dimensions?

• A. In infinite dimensions, Zorn's lemma ensures the extended functional remains linear — linearity fails automatically without it
• B. Finite-dimensional proofs assign values on a finite basis and stop; in infinite dimensions the extension must proceed 'one dimension at a time' through potentially uncountably many steps, and Zorn's lemma provides the maximality argument asserting the process terminates at a full extension
• C. Zorn's lemma is needed because the dual space of any infinite-dimensional normed space is empty without an axiom of choice argument
• D. The norm-preserving inequality |Φ(x)| ≤ ‖φ‖·‖x‖ requires well-ordering in infinite dimensions to establish

Answer: (B)

In finite dimensions, you extend a functional by freely choosing values on additional basis vectors — a finite number of choices suffice. In infinite dimensions, a subspace can have infinite codimension, requiring infinitely many extension steps. You can show that any partial extension (defined on some intermediate subspace) can be extended by one more dimension while preserving the norm bound. But asserting that this process reaches a complete extension defined on all of X requires Zorn's lemma: among all norm-preserving extensions defined on various subspaces, there is a maximal one — and maximality implies it is defined on all of X.

Question 3

The geometric form of Hahn-Banach — that a convex set and a point outside it can be separated by a hyperplane — is the infinite-dimensional generalization of a fact that is visually obvious in finite-dimensional spaces.

• T. True
• F. False

Answer: True

In ℝ² or ℝ³, it is geometrically clear that a convex set (a disk, a convex polygon) and an external point can be separated by a line or plane tangent to the set. This 'supporting hyperplane' intuition generalizes directly to infinite-dimensional Banach spaces via the geometric Hahn-Banach theorem. The difficulty is not conceptual but technical: in infinite dimensions, 'hyperplane' means the kernel of a bounded linear functional, and existence cannot be assumed without proof. The geometric and analytic forms of Hahn-Banach are two expressions of the same underlying fact about the richness of the dual.

Question 4

If the Hahn-Banach theorem failed to hold in a normed space X, the dual space X* might still contain enough non-trivial functionals to support weak convergence and reflexivity theory.

• T. True
• F. False

Answer: False

This is precisely what makes Hahn-Banach a cornerstone rather than a technical lemma. Without it, there is no guarantee that X* contains any non-zero bounded linear functional at all — the dual might be trivial. Weak convergence of a sequence (x_n → x weakly) is defined by φ(x_n) → φ(x) for all φ ∈ X*; if X* is trivial, every sequence weakly converges to everything and the concept collapses. Reflexivity (X ≅ X**) requires a rich double dual, which requires a rich dual. Point separation, weak topologies, adjoint operators — all depend on the dual being non-trivially large, which Hahn-Banach certifies.

Question 5

The Hahn-Banach theorem is called a 'cornerstone' of functional analysis. Explain why, focusing not on what the theorem says but on what would fail without it.

Model answer: Without Hahn-Banach, there is no guarantee that a normed space's dual X* contains any non-trivial bounded linear functionals — it might consist only of the zero functional. This would collapse the entire superstructure built on dual spaces. Weak convergence (x_n → x weakly iff φ(x_n) → φ(x) for all φ ∈ X*) becomes trivial or vacuous. Reflexivity (X ≅ X**) requires a rich double dual, which requires a rich dual. Adjoint operators are defined through the dual and would be meaningless. Point separation — the minimum condition for using duals to distinguish elements of X — would fail. Hahn-Banach is a cornerstone not because it provides a formula or algorithm, but because it certifies that the dual space is large enough to do mathematics with. The theorem's content is existential: such-and-such exists. Its importance is that without that existence guarantee, the foundations of Banach space theory — weak topologies, duality, reflexivity, the Hahn-Banach separation argument in convex analysis — would all be built on sand.

Contrast this with theorems like the open mapping theorem or closed graph theorem, which describe structural properties of operators. Hahn-Banach is more primitive: it asserts the richness of a fundamental object (the dual space) on which everything else depends. This is why it is proved early in functional analysis and used everywhere thereafter.