As the title says, I would like to know what the fundamental theorem of algebraic K-theory would say over the field with one element. Recall that the fundamental theorem of K-theory provides a decomposition $K_i(R[T,T^{-1}])\cong K_i(R)\oplus K_{i-1}(R)$ for $R$ regular.

Now turn to the philosophical part, the field $\mathbb{F}_1$. For a root system $\Phi$, the natural candidate for the algebraic group of type $\Phi$ over $\mathbb{F}_1$ is the Weyl group $W(\Phi)$ - the formula for the group order is right as noted by Tits, see also several other MO questions on $\mathbb{F}_1$ such as this or this. Now I want to understand the loop group of this group over $\mathbb{F}_1$. The notation would probably be something like $G(\Phi,\mathbb{F}_1[T,T^{-1}])$ and I apologize if anyone is offended by this abuse of notation. I would expect the affine Weyl group for the affine root system $\tilde{\Phi}$ to be a reasonable candidate.

The above philosophical guessing has some consequences, and I would like to know if the following statements have been considered, and if they are true or false. In the following, I specialize to the case of root systems $A_n$ and $\tilde{A}_n$, just for definiteness. In that case $W(A_n)\cong S_{n+1}$ is the symmetric group, and the affine Weyl group $W(\tilde{A}_n)$ is an extension of $S_{n+1}$ by $\mathbb{Z}^{n}$, where $\mathbb{Z}^n$ is identified with the hyperplane $x_1+\cdots+x_{n+1}=0$ in $\mathbb{Z}^{n+1}$ with the permutation action of $S_{n+1}$. Now here are the questions:

  1. Are there stabilization theorems for the homology of $W(\tilde{A}_n)$? How do these stabilization theorems compare to those for $S_{n+1}$, in particular is the stable range the same or one less? More generally, has the homology of $W(\tilde{A}_n)$ been computed by any chance (this is loosely related to this MO-question on cohomology of the action of $S_n$ on $T_n$).

  2. Provided there are stabilization theorems, have people considered the group completion of $\bigsqcup_n BW(\tilde{A}_n)$ and compared it to the plus-construction space $BW(\tilde{A}_\infty)^+$? A stabilization theorem should imply that the two are weakly equivalent.

  3. Does the fundamental theorem of algebraic K-theory hold in this situation? In other words, can we express the homotopy groups of $BW(\tilde{A}_\infty)^+$ in terms of stable homotopy groups of spheres? Maybe $BW(\tilde{A}_\infty)^+$ even splits as two copies of the sphere spectrum (one copy suitably shifted)?

Of course, similar questions can be formulated for the other classical series, and I would be happy to know the answers in these cases as well.

  • $\begingroup$ I don't understand your description of $W(\tilde{A}_n)$. How is $S_{n+1}$ supposed to act on $\mathbb{Z}^n$? $\endgroup$ – Oscar Randal-Williams May 17 '14 at 19:13
  • 2
    $\begingroup$ Sorry, I edited the question to clarify the action. It is the action $S_{n+1}$ on a hyperplane in $\mathbb{Z}^{n+1}$. $\endgroup$ – Matthias Wendt May 17 '14 at 19:31

Let $G_n := W(\tilde{A}_{n-1})$. If I understand your description correctly, there is an extension $$1 \to G_n \to S_{n} \wr \mathbb{Z} \overset{sum}\to \mathbb{Z} \to 1$$ and so a $\mathbb{Z}$-Galois cover $BG_n \to B(S_{n} \wr \mathbb{Z})$. There are exterior products $$\mu_{n,m} : G_n \times G_m \to G_{n+m}$$ given by concatenation, and these make $M := \coprod_{n \geq 0} BG_n$ into a homotopy commutative topological monoid. There are similar maps making $M' := \coprod_{n \geq 0} B(S_{n} \wr \mathbb{Z})$ into a homotopy commutative topologial monoid,

By the group-completion theorem of McDuff--Segal, the associated maps $$\mathbb{Z} \times BG_\infty \longrightarrow \Omega B M \quad \quad \mathbb{Z} \times B(S_\infty \wr \mathbb{Z}) \longrightarrow \Omega B M'$$ are both homology equivalences. In fact, by an addendum to the McDuff--Segal work in my paper "Group-Completion", local coefficient systems, and perfection, Quarterly Journal of Mathematics 64 (3) (2013) 795-803, these maps are both in fact acyclic, so yield homotopy equivalences $$\mathbb{Z} \times BG_\infty^+ \simeq\Omega B M \quad \quad \mathbb{Z} \times B(S_\infty \wr \mathbb{Z})^+ \simeq \Omega B M'.$$

The space $B(S_n \wr \mathbb{Z})$ can be modelled as configurations of $n$ points in $\mathbb{R}^\infty$ with labels in $B\mathbb{Z} = S^1$. As such, the group completion $\Omega B M'$ can be approached using Segal's "scanning" technique, and there is a homotopy equivalence $$\Omega BM' \simeq Q(B\mathbb{Z}_+)$$ to the infinite loop space of the spectrum $\mathbf{S}^0 \vee \mathbf{S}^1$.

To get at $\Omega BM$, we can use that the fibration sequence $$BG_\infty \to B(S_\infty \wr \mathbb{Z}) \to B\mathbb{Z}$$ satisfies a condition (e.g. Berrick's) to be plus-constructible, so there is a fibration sequence $$\Omega BM \to Q(B\mathbb{Z}_+) \to B\mathbb{Z}.$$ Thus $\Omega BM$ is the infinite loop space of the spectrum $\mathbf{S}^0 \vee \overline{\mathbf{S}}^1$, where $\overline{\mathbf{S}}^1$ is the 1-connected cover of $\mathbf{S}^1$.

I have carefully avoided discussion of homological stability for the groups $G_n$ (it is not, despite what the question says, necessary (or sufficient) to compare $\mathbb{Z} \times BG_\infty^+$ with $\Omega BM$). I think that homological stability does hold, but it requires a technique not yet in the literature. This technique will hopefully appear in a forthcoming paper.

  • $\begingroup$ Perfect. This is exactly the answer I was hoping for. So it seems that the fundamental theorem of K-theory does indeed hold. More in the $\mathbb{A}^1$-homotopy direction: I guess we also see from this that $S^1$-loop spaces and $\mathbb{G}_m$-loop spaces agree in this case - very nice that the configuration spaces from the recognition principle pop up. $\endgroup$ – Matthias Wendt May 18 '14 at 11:53
  • $\begingroup$ I'm not sure about the fundamental theorem holding: the calculation I gave shows that $K_1(\mathbb{F}_1[t, T^{-1}])=\mathbb{Z}/2$, which does not contain a summand of $K_0(\mathbb{F}_1)= \mathbb{Z}$. It does however hold in degrees above 1. Is this what the fundamental theorem should give? $\endgroup$ – Oscar Randal-Williams May 18 '14 at 19:01
  • $\begingroup$ Right. So the wreath product $S_n\wr \mathbb{Z}$ would be the better candidate for the $GL_n(\mathbb{F}_1[T,T^{-1}])$? One thing that is not yet clear to me is the difference between $SL_n$ and $GL_n$. These corresponding plus construction spaces have the same homotopy above degree 1 but differ in degree 1. Maybe that explains the difference, i.e. that $G_n$ is $SL_n$ and the wreath product is $GL_n$? $\endgroup$ – Matthias Wendt May 19 '14 at 8:32
  • 1
    $\begingroup$ Hi @MatthiasWendt. This question and its connection with A^1 homotopy is very interesting - could you elaborate more on your comment that "$S^1$-loop spaces and $\mathbb{G}_m$-loop spaces agreeing" in relation to the question? $\endgroup$ – Elden Elmanto Feb 9 '16 at 5:22

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.