I write the question for algebraic cobordism but I have the analogue question for classic cobordism.

The spectrum representing algebraic cobordism $$ \mathbf{MGL}=(*, \mathrm{Th}(1) , \ldots , \mathrm{Th}(n), \ldots) $$ is given by the Thom spaces $\mathrm{Th}(n)$ of the universal bundles of the Grassmanians $\mathrm{Gr}_n$. Let $\mathbf{E}=(E_n)_{n\in\mathbb{N}}$ be the spectrum of a cohomology theory with Chern classes (or oriented), it is very easy to write the universal map $\varphi^E\colon \mathbf{MGL}\to \mathbf{E}$ since it is the morphism of spectra having at the $n$-th level the Thom class $\mathit{th}^\mathbf{E}(\mathrm{Th}(n)) :\mathrm{Th}(n)\to E_n$.

Now, for the concrete case of motivic cohomology (or singular cohomology in the topological case) one can check that at the (0,0)-level the map $$ \mathbf{MGL}^{0,0}(X)\to H^{0}(X,\mathbb{Z}(0)) $$ induced by $\varphi^\mathbf{H}$ "is" the rank function. More concretely, the above arrow is the composition of the following: $$ \mathbf{MGL}^{0,0}(X)\xrightarrow{\varphi^K} K_{0}(X)\xrightarrow{\mathrm{rank}}H^{0}(X,\mathbb{Z}(0)) $$ One can check this explicitly, for example, whenever you know that $\mathbf{MGL}^{0,0}(X)$ is generated by morphisms $f\colon Y\to X$ where $\mathrm{dim} Y=\mathrm{dim}X$. This suggests that the unstable map $$ (\varphi^H)_0: \Omega^\infty \mathbf{MGL}\to K(0,0) $$ defined as the adjoint of the composition in $\mathbf{SH}$ of $\Sigma^\infty \Omega^\infty \mathbf{MGL}\to \mathbf{MGL}\xrightarrow {\varphi^H} \mathbf{H}$, should equal the composition in the unstable homotopy category of $$ \Omega^\infty \mathbf{MGL} \xrightarrow{(\varphi^K)_0}\mathbb{Z}\times \mathrm{Gr}\xrightarrow{\mathrm{rank}} K(0,0), $$ where $(\varphi^K)_0$ is the defined analogously to $(\varphi^H)_0$ and $K(0,0)$ is the space at level zero of the spectrum $\mathbf{H}$ representing motivic cohomology. To sum up, my question is:

Does $(\varphi^H)_0$ equal to $\mathrm{rank}\circ (\varphi^K)_0$?

On top of the motivation I gave above and to rephrase the question. It is already known that, over a field, $\Omega^\infty \mathbf{MGL}=\mathbb{Z}\times\mathrm{Hilb}_\infty^{\mathrm{lci}}(\mathbb{A}^\infty)^+$ (see [1]) and the rank function on $K$-theory is given by the projection towards the first factor of $\mathbb{Z}\times \mathrm{Gr}$, so it is very likely that $(\varphi^H)_0$ is also the projection towards the first factor composed with its natural map towards $K(0,0)$.

If you know any reference of a computation similar to this for classic cobordism I would also thank that.

  • 1
    $\begingroup$ I am not sure I understand your question, but does section 9 of this paper help (in particular corollary 9.2)? We wrote it to give a neat picture of the current state of affairs of geometric models for motivic cohomology theories $\endgroup$ May 27, 2021 at 15:04

1 Answer 1


Let me first write what happens for classical cobordism. You are basically asking whether the map $\operatorname{MU}\to H\mathbb{Z}$ factors through the projection $\operatorname{ku}\to H\mathbb{Z}$. But this is clear, since all spectra in sight are connective and we have an equivalence $$\operatorname{Map}(E,H\mathbb{Z})\cong \operatorname{Map}(\pi_0E,\mathbb{Z})$$ for every connective spectrum $E$ given by the fact that $\pi_0$ is the left adjoint of the inclusion of discrete spectra (i.e. abelian groups) into connective spectra. This adjunction by the way works both in spectra and abelian groups and $E_\infty$-ring spectra and commutative rings, in which case it's clear that the right hand side is just a point both for $\operatorname{MU}$ and $\operatorname{ku}$.

Something similar happens in motivic cohomology. Here the statement is that the map $\operatorname{MGL}\to H\mathbb{Z}$ factors uniquely through the orientation $\operatorname{MGL}\to\operatorname{kgl}$ (here with $\operatorname{kgl}$ we mean the very effective cover of the motivic spectrum $\operatorname{KGL}$). Instead of taking homotopy groups the correct thing to do is taking slices. Indeed the both map $\operatorname{MGL}\to H\mathbb{Z}$ and $\operatorname{kgl}\to H\mathbb{Z}$ exhibit $H\mathbb{Z}$ as the zero slice of the respective spectrum and $\operatorname{MGL}\to\operatorname{kgl}$ induces an equivalence on 0-slices.

  • $\begingroup$ Thank you very much, Denis. That indeed is a nice solution. One more thing, do you know a reference where I can read about the adjunction between groups/rings and spectra/ring spectra? The references I see only do the construction of $H$ and I am interested in reading in detail the adjoint map $1\to H\pi_0$, I don't see how to do it. Thanks in advance $\endgroup$
    – Tintin
    Jun 3, 2021 at 13:37
  • 1
    $\begingroup$ @Tintin For spectra this is just the standard axioms of a t-structure (the functor is the 0-truncation functor from connective spectra to the heart). For ring spectra, a reference is Higher Algebra,, but it's probably simpler to prove by identifying connective $E_\infty$-ring spectra with models for a certain Lawvere theory, where you can just apply the corresponding statement for spaces. $\endgroup$ Jun 3, 2021 at 20:24

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.