In ABGHR Thom spectra are described in the following way: we start with a morphism of Kan complexes $X\to \mathbb{S}\text{-line}$, where $\mathbb{S}\text{-line}$ is an $\infty$-groupoid which is equivalent to the well known $BGL_1(\mathbb{S})$. Then the Thom spectrum $Mf$ is defined to be the colimit of the composition of functors (of $\infty$-categories) $X\to\mathbb{S}\text{-line}\to\mathbb{S}Mod$. If we have a commutative ring spectrum $R$ there is a functor $\mathbb{S}\text{-line}\to R\text{-line}$ and it is a theorem that the colimit of the composition $X\to\mathbb{S}\text{-line}\to R\text{-line}\to RMod$ is equivalent to $Mf\wedge R$. There is a fibration of $\infty$-groupoids $R\text{-triv}\to R\text{-line}$ that we should think of as being our model of $EGL_1(R)\to BGL_1(R)$. If the composition $X\to R\text{-line}$ lifts to $R\text{-triv}$ then we say that $Mf$ is $R$-oriented and we have a Thom isomorphism $Mf\wedge R\simeq X_+\wedge R$.

My question is the following: we know classically that $Mf\wedge Mf\simeq X_+\wedge Mf $ (perhaps there are some conditions I need to put on $f$ to make this true?). So, if we assume $X$ is an infinite loop space and $X\to \mathbb{S}\text{-line}$ is a morphism of infinite loop spaces then we should have that $X\to\mathbb{S}\text{-line}\to Mf\text{-line}$ lifts to $Mf\text{-triv}$, yielding the classical Thom isomorphism described above. Is there an easy way to see that there is such a lift in this framework?

A more general question is: we know that such lifts always give such Thom isomorphisms, but are there conditions under which we can deduce a lift from such an isomorphism?


Let's try to see what a lift $\tilde f : X\to R\textrm{-triv}$ is. Recall that $R\textrm{-triv}$ is the $\infty$-groupoid of $R$-lines with a specified isomorphism with $R$, so a lift $\tilde f$ corresponds to giving a natural equivalence of $f$ with the constant functor $X\to R\textrm{-line}$ with value $R$. This is the same as giving an $R$-module map from the colimit of $f$ to $R$ which is an equivalence after precomposing with all the inclusions $f(x)\to \textrm{colim }f$. But $\textrm{colim }f=Mf\wedge R$, so a lift is effectively giving an $R$-linear map $Mf\wedge R \to R$ such that for all $x$ the composition $f(x)\wedge R\to R$ is an equivalence. Moreover an $R$-module map $Mf\wedge R\to R$ is the same thing as a spectrum map $Mf\to R$.

That is, a lift is the same datum as a Thom class.

Specializing in your case it is easy to see that the identity $Mf\to Mf$ gives exactly the Thom class you need (because the map $f$ is multiplicative, that is a loop space map), so you're done.

  • $\begingroup$ Does that spectrum map you describe, $Mf\to R$ need to be multiplicative? $\endgroup$ – Jonathan Beardsley Nov 30 '14 at 17:18
  • 1
    $\begingroup$ No, why should it be? You certainly do not need it in the proof of the Thom isomorphism. In fact the Thom isomorphism is true even when $f$ is not a loop map (so $Mf$ has no ring structure) $\endgroup$ – Denis Nardin Nov 30 '14 at 17:20

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.