on the dimension of a C*-algebra

88 Views Asked by At

I was reading a book and came across the following: If $f$ is a linear functional on a C*-algebra $A$ such that $a\leq f(a)1$ for all $a\geq 0$, then $A$ must be finite dimensional.

One possible way to see this is by a way of contradiction. That is, if $A$ is infinite dimensional, the there must exist a linear functional $f$ such that the above inequality does not hold. I could not come up with any such example or could not see any other way to go about it. Any help would be highly appreciated.

1

There are 1 best solutions below

1
On BEST ANSWER

This is definitely using far heavier machinery than necessary. But assume to the contrary that we have such an $f$ defined on an infinite-dimensional $A$. Note that $f$ is clearly positive and therefore necessarily bounded. By passing to the double dual, we may assume $A$ is a von Neumann algebra and $f$ is normal. Any infinite-dimensional von Neumann algebra admits a separable infinite-dimensional abelian von Neumann subalgebra, so by restricting to such a subalgebra, we may assume $A$ is a separable infinite-dimensional abelian von Neumann algebra and $f$ is normal. But the structure of such an algebra is known and we see that there exists a sequence of nonzero projections $p_n$ converging to $0$ in SOT. But then $f(p_n)1 \to 0$ in norm since $f$ is normal, but $p_n$ all have operator norm $1$, contradicting the assumption that $p_n \leq f(p_n)1$ for all $n$.