6
$\begingroup$

This is a problem in exercise one of Murphy's book

Find an example of a nonabelian unital Banach algebra $A$, where the only closed ideals are $\{0\}$ and $A$.

But does such an algebra exist at all?

My argument is the following:

Let $a$ be an arbitrary nonzero element in $A$, if $a$ is not invertible, then $a$ is contained in a maximal ideal, which is closed. However, there is no such ideal thus every nonzero element must be invertible. Then Gelfand-Mazur says $A$ is the complex numbers and thus must be abelian.

What is the problem with my argument?

Thanks!

$\endgroup$
4
$\begingroup$

The problem is that not being invertible doesn't imply being contained in a maximal two-sided ideal if $A$ is noncommutative. For example, there are plenty of elements of $\mathcal{M}_n(\mathbb{C})$ that are not invertible but since this algebra is simple (which incidentally implies that it's already an answer to your question when $n \ge 2$) all nonzero elements generate the unit two-sided ideal.

What is true is that not being left-invertible is equivalent to being contained in a maximal left ideal and not being right-invertible is equivalent to being contained in a maximal right ideal.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.