In a non-reflexive Banach space X, what is the essential difference between weak convergence and weak* convergence in X*?
AWeak* convergence is stronger — it requires fₙ(x) → f(x) for more test functions
BWeak* convergence tests functionals against elements of X; weak convergence in X* tests against all elements of X** (the double dual)
CThey are equivalent by the Hahn-Banach theorem for all Banach spaces
DWeak* convergence is defined only for Hilbert spaces, not general Banach spaces
Weak* convergence in X* requires fₙ(x) → f(x) for every x ∈ X — testing against the original space. Weak convergence in X* would require testing against every element of (X*)* = X**, the double dual. Since X embeds isometrically into X** but may not equal it (when X is not reflexive), weak* convergence imposes fewer conditions. It is a coarser topology: more sequences converge weak* than weakly. In a reflexive space X = X** and the two coincide.
Question 2 Multiple Choice
A student claims: 'Banach-Alaoglu is unsurprising — any bounded sequence in a Banach space has a convergent subsequence.' What is the critical error in this reasoning?
ABounded sequences only have convergent subsequences in finite-dimensional spaces; Riesz's theorem shows the unit ball is not norm-compact in infinite dimensions
BAlaoglu's theorem applies to weak convergence, not weak* convergence
CBounded sequences always converge in norm if the space is complete
DThe result only holds when the dual space X* is separable
Riesz's theorem states that the closed unit ball in an infinite-dimensional normed space is never compact in the norm topology — there always exist sequences with no norm-convergent subsequence. Alaoglu rescues compactness by switching topologies: the closed unit ball in X* IS compact in the weak* topology (a coarser topology, so compactness is easier to achieve). The student's claim confuses the weak* topology with the norm topology. This is why Alaoglu is profound, not trivial.
Question 3 True / False
In a reflexive Banach space, every weakly convergent sequence in X* is also weak* convergent.
TTrue
FFalse
Answer: True
In a reflexive space, X is isometrically isomorphic to X** via the canonical embedding — every bounded linear functional on X* is represented by an element of X. Therefore the weak topology on X* (testing against X**) and the weak* topology on X* (testing against X) coincide. Any weakly convergent sequence is automatically weak* convergent and vice versa.
Question 4 True / False
If (fₙ) converges weak* to f in X*, then (fₙ) converges in norm to f.
TTrue
FFalse
Answer: False
Weak* convergence is strictly weaker than norm convergence. Norm convergence requires ‖fₙ − f‖ → 0, meaning the functionals become uniformly close on the entire unit ball. Weak* convergence only requires fₙ(x) → f(x) pointwise for each fixed x ∈ X — there is no uniformity requirement. A sequence can converge weak* while ‖fₙ − f‖ remains bounded away from zero. The Banach-Alaoglu theorem exploits exactly this gap.
Question 5 Short Answer
Why is the weak* topology — rather than the norm topology — the right setting for Alaoglu's compactness theorem, and what makes this theorem useful in practice?
Think about your answer, then reveal below.
Model answer: Riesz's theorem shows the unit ball in an infinite-dimensional space is never norm-compact, so norm topology cannot supply compactness. The weak* topology on X* is defined only by pointwise convergence against elements of X — a coarser topology with fewer open sets — making it easier for sequences to converge and for sets to be compact. Alaoglu's theorem states the closed unit ball in X* is weak* compact. In practice, this lets you take a bounded sequence of approximate solutions (functionals or measures), extract a weak* convergent subnet via Alaoglu, and show the weak* limit is an exact solution — a pattern that drives existence proofs across analysis, PDEs, and variational calculus.