Let's say that in a unital C* algebra, we have $b \geq a \geq 0$ and $a$ is invertible. Then $b$ is also invertible. Can we conclude that $a^{-1} \geq b^{-1}$? If so, why? Can any related statement be made if we only assume $a \leq b$? (And maybe that $a$, and hence $b$ are self adjoint.)


Yes. If $b \ge a \ge 0$ and $a$ is invertible, then $a^{-1/2} b a^{-1/2} \ge a^{-1/2} a a^{-1/2} = 1$ so $ a^{1/2} b^{-1} a^{1/2} = (a^{-1/2} b a^{-1/2})^{-1} \le 1$ and then $b^{-1} = a^{-1/2} a^{1/2} b^{-1} a^{1/2} a^{-1/2} \le a^{-1}$. Without the $\ge 0$ it's false, e.g. try $a = -1$ and $b = 1$.

  • $\begingroup$ actually I saw that argument but I just don't understand why the special case you use in there holds. You're saying that if $a \leq 1$ and $a$ is invertible, then $a^{-1} \geq 1$. (and vice versa) $\endgroup$ – Jeff Aug 17 '12 at 7:35
  • 1
    $\begingroup$ Oh wait now I finally see. The way to see this fact is to use the functional calculus, and compare $t$, $1/t$, and the constant function $1$ defined on the spectrum which is a compact subset of $(0, 1]$. $\endgroup$ – Jeff Aug 17 '12 at 7:38
  • $\begingroup$ @Jeff: You can also show that directly, just using the spectrum, if you defined $b ≥ a$ by $\mathrm{spec}(b-a) \subseteq [0, ∞)$, which I believe is how it is commonly done for C*-algebras. $\endgroup$ – bodo Jul 3 '15 at 8:14

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.