10
$\begingroup$

So much is known about hermitian and non-hermitian idempotents in a $C^*$-algebra, that someone must have written down the following.

Theorem The universal $C^*$-algebra generated by one element $x$ subject to the relations $x^2=x$ and $ \| x \| \leq C$ is isomorphic to $$ \mathcal{I}_D= \left\{ f\in C\left([0,D],\mathbf{M}_{2}(\mathbb{C})\right) \left| f(0)\in\left[\begin{array}{cc} \mathbb{C} & 0\\ 0 & 0 \end{array}\right] \right. \right\} $$ where $D = \sqrt{C-1}$ and the isomorphism is specified by $$ x \mapsto \left[\begin{array}{cc} 1 & 0\\ t & 0 \end{array}\right], $$

This is not hard to prove, and makes it easy to see why an idempotent is homotopic to a projection.

Is there a reference for this result, that explicitly mentions a universal $C^*$-algebra or that spells out the universal property?

$\endgroup$

1 Answer 1

6
$\begingroup$

I almost found it: Example 4.4 in "Presentations and Tietze transformations of C*-algebras" by Will Grilliette, New York J. Math. 18 (2012) 121--137. The generator in the concrete algebra is not explicit, and must not be $$ \left[\begin{array}{cc} 1 & 0\\ t & 0 \end{array}\right] $$ as the length of the interval does not come out the same.

I am still interested in any earlier reference that might exist.

$\endgroup$
3
  • 3
    $\begingroup$ Shouldn't I lose point here, for having an intermittent memory and answering my own question? $\endgroup$ Dec 20, 2013 at 16:31
  • $\begingroup$ Did you ever find an earlier reference than this? Or would you think it is reasonable to call this a “folklore” result? $\endgroup$ Feb 20, 2023 at 21:11
  • 1
    $\begingroup$ @Chris Ramsey I never found an older reference. I think calling this folklore is fair. $\endgroup$ Feb 20, 2023 at 22:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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