2
$\begingroup$

Suppose that $A$ is a Banach algebra with unit. We can consider $GL(A)$ as a topological group in either the discrete topology or the topology that it inherits from the norm topology of $A$, and the identity map $i$ determines a continuous map $Bi: BGL(A) \longrightarrow B^{top}GL(A)$. If we apply Quillen's plus construction, we then obtain a continuous map $(Bi)^+: BGL(A)^+ \longrightarrow B^{top}GL(A)^+$. Because $\pi_1(B^{top}GL(A))$ is abelian, we know that $B^{top}GL(A)^+$ is (homotopy equivalent to) $B^{top}GL(A)$. Therefore we have an induced homomorphism $(Bi)_1^+: \pi_1(BGL(A)^+) \longrightarrow \pi_1(B^{top}GL(A))$. Of course, we know that $\pi_1(BGL(A)^+)$ is isomorphic to $K_1^{alg}(A) = GL(A)/E(A)$, while $\pi_1(B^{top}GL(A))$ is the topological (or operator algebraic) group $K_1^{top}(A) = GL(A)/GL(A)_0$. Now, there is an obvious map from $K_1^{alg}(A)$ to $K_1^{top}(A)$; namely, the quotient map, but it not obvious (to me) that $(Bi)_1^+$ is that map. Furthermore, in Loday's discussion of relative Banach algebra $K$-theory in his book "Cyclic Homology" (Section 11.5.2, p. 372), he made an offhand comment that suggests that $(Bi)_1^+$ can have a nontrivial cokernel in a group $K_0^{rel}(A)$ (that he did not define). Can anyone tell me what is known about this map $(Bi)_1^+$?

$\endgroup$
0

1 Answer 1

3
$\begingroup$

The $(Bi)^+_1$ is obviously surjective, since it is induced by the identity on $GL(R)$. The only reasonable definition for $K_0^{rel}(A)$ for a Banach algebra is the zero abelian group, since also $K^{alg}_0(A)=K^{top}_0(A)$.

$\endgroup$
3
  • $\begingroup$ Could you elaborate, please? $\endgroup$
    – Efton Park
    Mar 19, 2015 at 14:47
  • 1
    $\begingroup$ It is hard to say more. Any map $G \to H$ induces a map $BG \to BH$ with $\pi_1(BG) \to \pi_1(BH)$ being induced by $G \to H$. The same also if you perform the $+$-construction. Thus, $K_1^{alg}(A) \to K_1^{top}(A)$ is always surjective. Since, $K_0$ makes no distinction between "algebraic" and "topological", the only choice for $K_0^{rel}(A)$ is the zero abelian group. I think, that Loday did not mean to say that $K_0^{rel}(A)\neq 0$ in his comment - only that one could define it somehow, namely as the zero group. $\endgroup$ Mar 19, 2015 at 21:58
  • $\begingroup$ It was Loday's comment that has always made me wonder about this. It is sad that he is not with us so that we could ask him about it. $\endgroup$
    – Efton Park
    Mar 20, 2015 at 15:46

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.