This is a very general question-- I'm really after any references to anything similar...

Let $A$ be a $C^*$-algebra equipped with a continuous one-parameter group of automorphisms $(\alpha_t)_{t\in\mathbb R}$ and let $\pi:A\rightarrow B(H)$ be a (non-degenerate) *-representation. I'm after conditions which give a $\sigma$-weakly continuous one-parameter group of automorphisms of $B(H)$, say $(\beta_t)$, such that $\pi\circ\alpha_t = \beta_t\circ\pi$.

As $\beta_t$ will be implemented as $\beta_t(x) = U_t x U_t^*$ where $(U_t)$ is a continuous group of unitaries (that is, a continuous unitary representation of $\mathbb R$). So I'm exactly asking:

When does $\pi$ arise as part of a covariant representation of $(A,\alpha)$?

As covariant representations biject with representations of the crossed product $A \rtimes_\alpha \mathbb R$, it seems that in principle knowing $A \rtimes_\alpha \mathbb R$ should be enough. But this seems terribly abstract!


Borchers, H.-J. On the implementability of automorphism groups. Comm. Math. Phys. 14 1969 305–314. MathSciNet Project Euclid

Borchers defines $\pi$ to be covariant extendible if $\pi$ is (unitarily) equivalent to a subrepresentation of a covariant representation (as it happens, this is enough for my application). Then we have

Theorem: $\pi:A\rightarrow B(H)$ is covariant extendible if and only if for each $\xi,\eta\in H$ the vector state $\mu:a\mapsto (\pi(a)\xi|\eta)$ is such that $\lim_{t\rightarrow 0} \|\mu\circ\alpha_t-\mu\|=0$.

Pedersen's book has lots more, it turns out.


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.