Let $A$ be a finite dimensional $\mathrm{C}^*$-algebra $$A=\bigoplus_{j=1}^mM_{n_j}(\mathbb{C}).$$

Question 1: Is it true that every state on $A$ is of the form: $$\rho=\bigoplus_{j=1}^m \rho_{j},$$

where each $\rho_j$ is a positive functional on $M_{n_j}(\mathbb{C})$?

Isn't this obviously true... or am I missing something? More importantly for me:

Question 2: Moreover, either as a corollary or a separate result, is each pure state of $A$ equal to zero on every factor except for one? i.e. each pure state on $A$ arises from a pure state on one of the matrix factors.


For question 2, yes. For question 1, almost yes, every state is of the form $$\rho=\bigoplus^{m}_{j=1}\lambda_{j}\rho_{j}$$ where $\rho_{j}$ is a state on $M_{n_{j}}(\mathbb{C})$ and $\lambda_{j}\in\mathbb{R}_{\geq 0}$ such that $\sum^{m}_{j=1}\lambda_{j}=1$. This to ensure that $$\rho(I_{A})=\sum^{m}_{j=1}\lambda_{j}\rho_{j}(I_{n_{j}})=\sum^{m}_{j=1}\lambda_{j}\rho_{j}=1.$$

  • $\begingroup$ Thank you. In fairness I knew it was a convex combination so I wrote that $\rho_j$ were positive functionals. $\endgroup$ – JP McCarthy Jun 25 at 13:30
  • $\begingroup$ True, but taking arbitrary positive functionals would not automatically yield a state. $\endgroup$ – Floris Claassens Jun 25 at 13:33

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.