In his notes on elliptic cohomology, Lurie defines the multiplicative group $\mathbb{G}_m$ over a ring spectrum $A$ as $\operatorname{Spec} A[\mathbb{Z}]$. What is the value $\mathbb{G}_m(B)$ of the represented functor at an $A$-algebra $B$? If this is too hard to say in general: Are there at least any specific examples, other than Eilenberg-MacLane spectra, where one does know the answer?


It is slightly complicated.

One has a number of adjunctions: $$ \begin{eqnarray*} \mathbb{G}_m(B) &=& Hom_{A-alg}(A[\mathbb{Z}],B) \\ &\simeq& Hom_{E_\infty-rings}(\mathbb{S}[\mathbb{Z}],B)\\ &\simeq& Hom_{E_\infty-spaces}(\mathbb{Z},GL_1(B))\\ &\simeq& Hom_{spectra}(H\mathbb{Z},gl_1(B)). \end{eqnarray*} $$ (Note these adjunctions are weak equivalences of spaces, and the last two adjunctions require a fair amount of theory to make rigorous.)

The problem is that it is usually quite difficult to compute the maps out of the Eilenberg-Mac Lane spectrum $H\mathbb{Z}$ unless the target is also an Eilenberg-Mac Lane space. In the case where the algebra $B$ comes from a simplicial commutative ring, this is true and so one at least knows that the set of homotopy classes of maps $[H\mathbb{Z}, gl_1(B)]$ surjects onto $\pi_0(B)^\times$. Even for complex K-theory, the calculation is somewhat involved (but doable), but the only method that I can immediately think of involves the Bousfield-Kuhn functor.

  • $\begingroup$ I forgot an important special case of this. If $\pi_0(B)$ is a $\mathbb{Q}$-algebra, then $\mathbb{G}_m(B)$ is isomorphic to $\pi_0(B)^\times$ after passing to homotopy classes of maps. $\endgroup$ – Tyler Lawson Aug 13 '10 at 14:24
  • 1
    $\begingroup$ Tyler, can't we say that if $\pi_0(B)$ is a $\mathbb{Q}$-algebra, then $G_m(B)\simeq GL_1(B)$? $\endgroup$ – Charles Rezk Aug 13 '10 at 17:52
  • $\begingroup$ @Charles: Well, yes. I guess I should have decided once and for all whether I was talking about spaces of maps or homotopy classes of maps $\endgroup$ – Tyler Lawson Aug 13 '10 at 18:05
  • $\begingroup$ Tyler, Margolis shows it is not hard to compute maps out of $HF_p$: the Adams spectral sequence collapses to the 0-line because the Steenrod algebra is injective in the category of bounded below modules. From there, HZ is a Bockstein away. See ams.org/journals/proc/1974-043-02/S0002-9939-1974-0341488-9/… $\endgroup$ – Robert Bruner Apr 11 '15 at 18:11
  • $\begingroup$ Bob, that's a lovely result. (I suppose that things like this are predecessors to the Sullivan conjecture.) From that point I guess one needs tools for computing the cohomology of $gl_1(B)$, which may pick up some nontrivial Steenrod algebra summands where $B$ had none. $\endgroup$ – Tyler Lawson Apr 11 '15 at 20:06

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.