3
$\begingroup$

Consider the (non-unital) $\mathbb{C}$-algebra (point-wise multiplication) of $\mathcal{S}$ of Schwartz functions on $\mathbb{R}$.

Question: Does there exist some character (non-zero multiplicative functional to $\mathbb{C}$) $\omega$ of $\mathcal{S}$ that is not the evaluation map at any point in $\mathbb{R}$, i.e. does $\mathcal{S}$ as an algebra, admit characters other than the functionals given by Dirac measures?

Some remarks:

  1. This can not happen if we require $\omega$ to be continuous with respect to the $L^\infty$ norm (one can extend $\omega$ to $C_0(\mathbb{R})$, which is a commutative $C^\ast$-algebra, then use Gelfand-Naimark), my question does not pose any continuity restriction on $\omega$, and can be seen of an algebraic nature.

  2. This amounts to the question whether the ideal $C_c^\infty(\mathbb{R})$ of compactly supported smooth functions is contained in a codimension $1$ ideal of $\mathcal{S}$.

$\endgroup$
1
  • $\begingroup$ For reference, this old math.SE points to a paper that deals with the case of $C_0(X)$. $\endgroup$ Commented Jan 30, 2023 at 19:19

1 Answer 1

5
$\begingroup$

Let $m$ be a multiplicative functional. Let $A={\mathbb C}\oplus\mathcal S$ be the algebra $\mathcal S$ extended by the constant functions. This algebra is unital. Setting $m(f+\lambda)=m(f)+\lambda$ extends $m$ to a multiplicative functional on $A$. For $f\in A$ let $s(f)=\sup_{x\in\mathbb R}|f(x)|$. Assume $s(f)<1$. Then we claim that $\frac1{1-f}$ lies in $A$. We have $\frac1{1-f}-1=\frac f{1-f}$ and any derivative of the latter is of a quotient of a polynomial in the derivatives of $f$ divided by a power of $1-f$. One concludes that $\frac f{1-f}$ lies in $\mathcal S$, hence $1-f$ is invertible in $A$. For $\lambda\ne 0$ we have $f-\lambda=\lambda(\frac f\lambda -1)$, so if $|\lambda|>s(f)$, the latter is invertible. Now we have $m(f-m(f))=0$, hence $f-m(f)$ is not invertible, hence $|m(f)|\le s(f)$. This means that $m$ is continuous in the sup-norm. Hence it extends to a continuous functional on $C_c({\mathbb R})$, i.e., a Radon measure. For this to be multiplicative, it must be a point measure.

$\endgroup$

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.