It seems to be a well known fact that algebraic K-theory can be used to understand intersection theory, at least for varieties (or stacks!) over a field. A first glimpse of this result seems to be Bloch's formula (Quillen, 1972): $$ \operatorname H^p(X, \mathcal{K}_p) \cong \operatorname{CH}^p (X) $$ where $\mathcal{K}_p$ is the Zariski-sheafification of $K$-theory. This even induces an isomorphism of the respective graded algebras!

But as beautiful as this formula is, I've got a question:

can we actually extract information about the chow groups and its intersection products using this formula?

Of course, $K$-theory is notoriously complicated, but sheafifying it might make things more amenable (I think $\mathcal{K}_p = \underline{\mathbf{Z}}$ and $\mathcal{K}_1 = \mathcal{O}_X^\times$).

One example on how this might help, it seems that (Guillet, 1984) used the LHS to define an intersection theory for algebraic stacks, but it also seems to me (and I might be very wrong) that (Vistoli, 1989) approach bypassing this is what has become more prominent nowadays, and I don't know if Guillet's approach allows us to deduce anything that Vistoli's cannot.

On a similar note, Quillen says on his paper on higher algebraic K theory that the left hand side of Bloch's formula is "manifestly contraviariant in X"; which may be one advantage, although I don't quite see how to prove this (maybe it follows from earlier remarks - i didn't read the entire paper yet....). It seems that Quillen has convinced himself by this point that the LHS will be important for future work.

  • $\begingroup$ Just a remark on the "manifest" contravariance in $X$: it follows from the functoriality of cohomology because $\mathcal K_p$ is by definition a presheaf on all schemes. If one regards it as a presheaf of chain complexes in degree $0$ and Zariski-sheafify in the homotopical sense (i.e., take derived sections), one obtains a functor $R\Gamma(-,\mathcal K_p): Sch^{op} \to D(\mathbb Z)_{\leq 0}$, and one can then take cohomology groups. $\endgroup$ Commented Jul 19, 2021 at 10:27

1 Answer 1


Not an answer, but as you wrote "Of course, K-theory is notoriously complicated, but sheafifying it might make things more amenable", let me point out that Bloch's formula is also true for the Milnor $K$-theory sheaf instead of the Quillen $K$-theory sheaf. This does not help at all to compute any Chow group, but it makes it possible (and in easy examples possible in a concrete fashion) to write down an algebraic cycles as a representative in Cech cohomology of the Milnor $K$-theory sheaf. So this makes things even more amenable. Nonetheless, I am not sure this has any profound use. It is however definitely easier to work with such cycle representatives than trying to make a Cech representative of a Quillen K-theory sheaf explicit.


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.