Goldstine's theorem #
Goldstine's theorem says that the canonical embedding of a normed space E into its bidual
maps the closed unit ball of E onto a weak-star dense subset of the closed unit ball of the
bidual. We prove it for every closed ball: the weak-star closure of the image of
closedBall x r under NormedSpace.inclusionInDoubleDualWeak is the closed ball of radius r
about the image of x. In particular E is weak-star dense in its bidual.
No completeness or separability is assumed, and seminormed spaces are allowed. Together with
BanachβAlaoglu (WeakDual.isCompact_closedBall), Goldstine's theorem relates weak compactness
of the unit ball of E to reflexivity of E.
Main results #
NormedSpace.closure_image_inclusionInDoubleDualWeak_closedBall: Goldstine's theorem.NormedSpace.closure_image_inclusionInDoubleDualWeak_closedBall_zero_one: its unit-ball form.NormedSpace.denseRange_inclusionInDoubleDualWeak:Eis weak-star dense in its bidual.
References #
- R. E. Megginson, An Introduction to Banach Space Theory, Theorem 2.6.26.
Goldstine's theorem. The weak-star closure of the image of the closed ball closedBall x r
under the canonical embedding into the bidual is the closed ball of radius r about the image
of x.
Goldstine's theorem, unit-ball form: the closed unit ball of E has weak-star dense
image in the closed unit ball of its bidual.
A normed space is weak-star dense in its bidual.