3
$\begingroup$

Let $\mathfrak{A}$ be a C*-algebra, and let $\phi \in \mathfrak{A}^*$ be a self-adjoint bounded linear functional on $\mathfrak{A}$. Then there exists a unique pair $\phi^+, \phi^-$ of positive bounded linear functionals on $\mathfrak{A}$ such that $\phi = \phi^+ - \phi^-$ and $\| \phi \| = \left\| \phi^+ \right\| - \left\| \phi^- \right\|$.

My question is: If $\phi$ is tracial (i.e. $\phi(xy)=\phi(yx)$ for all $x, y \in \mathfrak{A}$), then are $\phi^+, \phi^-$ necessarily tracial?

$\endgroup$

2 Answers 2

4
$\begingroup$

Assuming $\mathfrak{A}$ is unital, every algebra element is a linear combination of unitaries, so a linear functional being tracial is equivalent to $\phi = \phi \circ \operatorname{Ad} u$ for every unitary $u$. In this case, for every unitary $u$ we have

$$ \phi = \phi \circ \operatorname{Ad} u = \phi^+ \circ \operatorname{Ad} u - \phi^- \circ \operatorname{Ad} u. $$

By the uniqueness of the decomposition you have $\phi^\pm = \phi^\pm \circ \operatorname{Ad} u$ and so $\phi^{\pm}$ are tracial.

$\endgroup$
0
$\begingroup$

By 1.14.3. Theorem. in "$C^*$-algebras and $W^*$-algebras" we have $\phi^+(x)=\phi(px)=\phi(xp)$ for a projection $p$ in $\mathfrak A^{**}$. Therefore $$\phi^+(xy)=\phi(pxy)=\phi(ypx)=\phi(pypx)=\phi^+(yx)$$

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .