5

Two commutative monoids $M,N$ have a tensor product $M\otimes N$ satisfying the universal property that there is a tensor-Hom adjunction for any other commutative monoid $L$: $$\text{Hom}(M\otimes N,L)\cong\text{Hom}(M,\text{Hom}(N,L)).$$ If $M$ and $N$ are abelian groups, then $M\otimes N$ agrees with the abelian group tensor product.

Now I would like to understand the derived tensor product and $\text{Tor}(M,N)$. Similar questions have been asked before, but I am interested in a very particular construction:

If we identify $M$, $N$, $\mathbb{N}$ as discrete commutative monoid spaces, we may take the relative smash product of $E_\infty$-spaces (see below) $M\wedge_{\mathbb{N}} N$. Each point $x\in M\otimes N$ corresponds to a connected component of $M\wedge_{\mathbb{N}} N$, and so we can take Tor groups based at x: $$\text{Tor}_n(M,N;x)=\pi_n(M\wedge_{\mathbb{N}} N;x).$$ If $M$ and $N$ are groups, the choice of basepoint is irrelevant (we can always shift to $x=0$), and we recover the usual Tor groups.

How might I go about computing this commutative monoid Tor? Here is a more specific question: if $\mathbb{N}\rightarrow R$ is a surjection of commutative semirings (such as $R=\mathbb{N}/(n=n+1)$), is there some $x\in R\otimes R\cong R$ such that $\text{Tor}(R,R;x)$ is nonzero? I suspect the answer is yes when $x=n$.

By the smash product of $E_\infty$-spaces, I mean for example $\infty$-categorical Day convolution of $\Gamma$-spaces, which satisfies the same tensor-Hom adjunction universal property. More generally, I would be interested in techniques to compute smash products of $E_\infty$-spaces, but I think this is very difficult, so I am asking about this toy example.

  • I don't think you can do much with $E_\infty$-spaces without homotopy invertibility. In the connected case this is a bonus anyway but you seem to need discrete monoids, and then I don't believe $E_\infty$-notions behave well enough to help with anything – მამუკა ჯიბლაძე Aug 10 '17 at 16:23
  • To put it another way, group completing commutative monoids is like passing to the stable category, so commutative monoids without homotopy inverses in some sense carry unstable homotopy information; then, I do not know any sense in which one could speak of $Е\infty$ such guys, as all $\infty$ gadgets presume everything around already stabilized – მამუკა ჯიბლაძე Aug 10 '17 at 16:32
  • Do we even know what's happening when R=N? Since you're working in E_infty land, N is no longer free. – Dylan Wilson Aug 10 '17 at 21:11
  • @DylanWilson Oops, that's because I made a mistake. I wanted to smash over the natural numbers. Otherwise, Tor(Z,Z) wouldn't be right either! I edited above. – John Berman Aug 10 '17 at 23:19
  • This is an interesting question. I believe that there is no reason why one couldn't compute things like the homology of the smash product of such things, at least theoretically. This would be hard though (if you want to smash over something other than N). You don't really need that the inputs are commutative monoids though do you? You just need that they are "modules" over N. – Sean Tilson Aug 15 '17 at 12:55

0 Answers0