In section 3, pg 8 of String theory on Elliptic curves they claim that for $X$ compact with $x_0$ an involution fixed point, the map $i:x_o\rightarrow X$ is equivariant and equivariantly split, which according to them implies

$KR^{j}(X) \approx KR^{j}(X - \{x_0 \}) \oplus KO^{j}(x_0)$

I don't understand what equivariantly split means here relative to the functor $KR^j( - )$ other than you can simply do this separation nor why this is so.

1) I would appreciate if someone could put in more steps into the isomorphism $KR^{j}(X) \approx KR^{j}(X - \{x_0 \}) \oplus KO^{j}(x_0)$

2) Is there a reference were they explain in more detail what is meant by equivariantly split? I haven't seen any mention of the term the way is being used here for any kind of equivariant cohomology theory

Note that it seems to be $KR$-theory with compact supports but I don't know if this is essential for the above argument.

Thank you for your time


Equivariantly split just means split in the category of spaces with a $C_2$-action, i.e. that there is an equivariant map $r:X→\{x_0\}$ such that $r\circ i$ is equivalent to the identity. This map is usually called a retraction of $i$. Of course the splitting in this case is trivial to construct.

Since $KR^j(-)$ is a functor from the category of spaces with a $C_2$-action to abelian groups, you have that the map $KO^j(*)=KR^j(\{x_0\})\leftarrow KR^j(X)$ is split too, but in the category of abelian groups this means precisely that if $KR^j(X,x_0):=\ker i^*$ we have $KR^j(X)\cong KR^j(X,x_0)\oplus KO^j(*)$.

To show that $KR^j(X,x_0)\cong KR^j(X\smallsetminus\{x_0\})$ in favorable cases, we are going to assume that $x_0$ has a filter of $C_2$-contractible neighborhoods $\{U_i\}$ (this is the case, e.g., if $X$ is a $C_2$-manifold). Then $KR$-cohomology with compact supports is, by definition, $$KR^j(X\smallsetminus\{x_0\})=\lim_i KR^j(X\smallsetminus U_i) ,.$$ Since by excision the map $KR^j(X\smallsetminus U_i)\cong KR^j(X,x_0)$ is an equivalence, we are done.


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.