Constructing a ring in Top by wedging and smashing pointed spaces

140 Views Asked by At

I'll list some things I believe I can do, any of which might I might not actually be able to, and then, if I'm right, I'll ask if there's any point to what I've done.

The category of pointed locally compact topological spaces is a symmetric monoidal category under the smash product $\wedge$. Its skeleton $T_*$ is a (large) semiring $(T_*,\vee,\wedge,\{0\},S^0)$ if we take the wedge sum to be $+$, the smash product to be $\cdot$, the point to be 0, and the 0-sphere to be 1.

The category of all topological spaces is also symmetric monoidal under the product $\times$, with coproduct $\amalg$. Taking $0 = \emptyset$ and $1 = \{0\}$, I can make another semiring $T$.

Fix a field $k$. Then graded $k$-vector spaces become a symmetric monoidal category under $\otimes_k$, and the skeleton $M$ of this category is a semiring $(M,\oplus,\otimes_k,0,k)$.

In general given a symmetric monoidal category with finite coproducts, if the monoidal product distributes over the coproduct, taking the skeleton gives a semiring.

The functor $\tilde H^*(-;k)$, reduced singular cohomology with field coefficients, is monoidal, and when we take the skeleton, becomes a semiring homomorphism $f \colon T \to M$.

The Grothendieck group functor $K\colon \mathrm{Mon} \to \mathrm{Grp}$ takes semirings to rings, because functors preserve diagrams, in particular for monoid objects. So I make my categories into rings $K(T_*)$, $K(T)$, and $K(M)$, the same way one does for the categories of vector bundles over a space or representations of a group.

I have a ring homomorphism $K(f)\colon K(T) \to K(M)$. (I would like to say I have something more, because of the induced pullback maps in cohomology, but I doubt they survive the $K$-equivalence relation.)

Does anything? What I've said seems like a natural thing to do, so I'm sure someone has thought of it before, but I imagine if it yielded anything informative, I'd have heard of it, so it probably turns out to be boring. What goes wrong? Why is this boring?

1

There are 1 best solutions below

7
On BEST ANSWER

The usual answer with these things seems to be the Eilenberg–Mazur swindle: for instance, take $Y$ to be the infinite wedge sum of any pointed space $X$, then $X \vee Y \cong Y$; so if you apply the Grothendieck group construction to the additive monoid, you get $X = 0$.

The moral of the story is that we cannot have a abelian group in which infinite sums make sense. So we are forced to restrict our attention to "finite" objects, as in e.g. the construction of $K_0 (A)$ for a ring $A$.