Chapter 2

Spaces

[6/100%]
Show
LaTeX
Problem 13

If M\mathbf{M} is a finite-dimensional linear manifold in a Hilbert space H\mathbf{H}, and if N\mathbf{N} is a subspace (a closed linear manifold) in H\mathbf{H}, then the vector sum M+N\mathbf{M} + \mathbf{N} is necessarily closed (and is therefore equal to the span M∨N\mathbf{M} \vee \mathbf{N}).

?
Problem 14

For which cardinal numbers mm is the lattice of subspaces of a Hilbert space of dimension mm modular? distributive?

?
Remark.
?

The collection of all subspaces of a Hilbert space is a lattice. This means that the collection is partially ordered (by inclusion), and that any two elements M\mathbf{M} and N\mathbf{N} of it have a least upper bound or supremum (namely the span M∨N\mathbf{M} \vee \mathbf{N}) and a greatest lower bound or infimum (namely the intersection M∩N\mathbf{M} \cap \mathbf{N}). A lattice is called distributive if (in the notation appropriate to subspaces)

L∩(M∨N)=(L∩M)∨(L∩N) \mathbf{L} \cap (\mathbf{M} \vee \mathbf{N}) = (\mathbf{L} \cap \mathbf{M}) \vee (\mathbf{L} \cap \mathbf{N})

identically in L\mathbf{L}, M\mathbf{M}, and N\mathbf{N}.

There is a weakening of this distributivity condition, called modularity; a lattice is called modular if the distributive law, as written above, holds at least when N⊂L\mathbf{N} \subset \mathbf{L}. In that case, of course, L∩N=N\mathbf{L} \cap \mathbf{N} = \mathbf{N}, and the identity becomes

L∩(M∨N)=(L∩M)∨N \mathbf{L} \cap (\mathbf{M} \vee \mathbf{N}) = (\mathbf{L} \cap \mathbf{M}) \vee \mathbf{N}

(with the proviso N⊂L\mathbf{N} \subset \mathbf{L} still in force).

Problem 15

For subspaces M\mathbf{M} and N\mathbf{N} of a Hilbert space, the vector sum M+N\mathbf{M} + \mathbf{N} is closed if and only if the modular equation

L∩(M∨N)=(L∩M)∨N \mathbf{L} \cap (\mathbf{M} \vee \mathbf{N}) = (\mathbf{L} \cap \mathbf{M}) \vee \mathbf{N}

is true whenever N⊂L\mathbf{N} \subset \mathbf{L}.

?
Problem 16

A Hilbert space is locally compact if and only if it is finite-dimensional.

?
Problem 17

A Hilbert space H\mathbf{H} is separable if and only if dim⁡H≤ℵ0\dim \mathbf{H} \leq \aleph_0.

?
Problem 18

For each measure in an infinite-dimensional Hilbert space, the measure of every non-empty ball is infinite.

?
Remark.
?

Here "measure" refers to a countably additive set function μ\mu defined on (at least) the collection of all Borel sets (the σ\sigma-field generated by the open sets), so that 0≤μ(M)≤∞0 \leq \mu (M) \leq \infty for all Borel sets MM, such that every non-empty open set has positive measure and μ\mu is invariant under translation (i.e., μ(f+M)=μ(M)\mu (f+M) = \mu (M) for every vector ff and for every Borel set MM).