Next: Dual spaces for Quotients
Up: Linear Functionals
Previous: Hahn-Banach Theorem
Consider
. Define
by
.
A moment's thought shows
is linear and
. Moreover, if
for all
, then
.
Theorem 10.2.1
There is a linear functional
such that
- If
, then
.
- If
with
, then
- If
then
(Not unique)
L is called the Banach Limit
Proof.
-
. Define
by
. This is a linear operator of norm 1.
Let
. As
is the kernel of
,
is a linear subspace.
Let
.
Claim:
.
As
,
. Let
. If for any
,
, then the
entry of
will be nonpositive. Then,
is at least
It remains only to consider
in
with
for each
. As
for some
for
, the
entry of
goes to 0. Thus,
.
By corollary (4), there is
with
and
and
.
Claim:
.
Let
. Consider
. This converges to
in
. Why? Let
. Pick
such that
,
. Then, if we ``prime"
or more times, we get a sequence with all entries at most
in absolute value. But,
. Since
,
proving the claim.
For
, if
, then
, as
.
It remains to prove (3). Suppose
,
for all
and
. WLOG,
. So,
for all
and
has entries between 0 and 1. So,
and
, which is a contradiction.
-
Every
can be written as
,
. Define
by
where
is the functional from case 1. A moments thought shows
is
-linear and properties 2,3, and 4 hold.
Clearly,
and
shows
.
As
, the simple functions in
are those of the form
where
are pairwise disjoint subsets of
. Then,
and
, so
on the simple functions. The simple functions are dense in an
space, so
.
Next: Dual spaces for Quotients
Up: Linear Functionals
Previous: Hahn-Banach Theorem
Brian Bockelman
2005-12-12